ed394172.pdf - eric - department of education

ED 394 172 AUTHOR TITLE REPORT NO PUB DATE NOTE AVAILABLE FROM PUB TYPE EDRS PRICE DESCRIPTORS ABSTRACT DOCUMENT RESUME EA 027 489 Leithwood, Kenneth; And Others Developing Expert Leadership for Future Schools. ISBNO-0-7507-0327-X 94 331p. Falmer Press, Taylor & Francis, Inc., 1900 Frost Road, Suite 101, Bristol, PA 19007 (paperback: ISBN-0-7507-0327-X; clothbound: ISBN-1-8500-0743-8). Books (010) Viewpoints (Opinion/Position Papers, Essays, etc.) (120) MF01/PC14 Plus Postage. Administrator Effectiveness; Educational Cooperation; Elementary Secondary Education; *Leadership; Leadership Qualities; Leadership Styles; Leadership Training; Management Development; *Problem Solving; *Professional Development; Teacher Administrator Relationship; Values This book describes change in the nature of school leadership as required by the nature of future schools. Two chapters in part 1 develop the perspective for leadership, based on problem solving, for the future. Chapter 1 argues that the problem for future leadership has three components: (1) developing a shared, defensible vision of a future school; (2) directly assisting school members in addressing problems encountered in achieving the vision; and (3) increasing the capacity of school members to address future challenges themselves. Chapter 2 describes what is known about the present state of school leadership. Seven chapters in part 2 describe the kind of leadership that will be productive for future schools. Chapter 3 raises issues involved in developing a shared school vision and develops one such vision. Chapters 4-6 offer a research-based model of leadership as a problem-solving process. Chapters 7-9 are devoted to selected aspects of problem solving in routine and nonroutine matters--the role of values, teacher development, and creation of a collaborative school culture. The six chapters comprising the third part address issues concerning the development of expert leadership. Chapter 10 assesses the value of a broad array of formal and informal experiences in developing expert leadership. Characteristics of formal preparation programs effective in developing expertise in routine and nonroutine matters are outlined in chapters 11 and 12. The next two chapters examine how administrator appraisal and selection processes might be designed and implemented. Finally, chapter 15 proposes five broad strategies that district leaders might use in their efforts to foster the development of leaders for future schools. Six figures, 15 tables, and an index are included. References accompany each chapter. The appendix contains the principal profile-based instrument. (LMI)

Upload: khangminh22

Post on 28-Feb-2023




0 download


ED 394 172






EA 027 489

Leithwood, Kenneth; And OthersDeveloping Expert Leadership for Future Schools.ISBNO-0-7507-0327-X94

331p.Falmer Press, Taylor & Francis, Inc., 1900 FrostRoad, Suite 101, Bristol, PA 19007 (paperback:ISBN-0-7507-0327-X; clothbound:ISBN-1-8500-0743-8).Books (010) Viewpoints (Opinion/Position Papers,Essays, etc.) (120)

MF01/PC14 Plus Postage.Administrator Effectiveness; Educational Cooperation;Elementary Secondary Education; *Leadership;Leadership Qualities; Leadership Styles; LeadershipTraining; Management Development; *Problem Solving;*Professional Development; Teacher AdministratorRelationship; Values

This book describes change in the nature of schoolleadership as required by the nature of future schools. Two chaptersin part 1 develop the perspective for leadership, based on problemsolving, for the future. Chapter 1 argues that the problem for futureleadership has three components: (1) developing a shared, defensiblevision of a future school; (2) directly assisting school members inaddressing problems encountered in achieving the vision; and (3)increasing the capacity of school members to address futurechallenges themselves. Chapter 2 describes what is known about thepresent state of school leadership. Seven chapters in part 2 describethe kind of leadership that will be productive for future schools.Chapter 3 raises issues involved in developing a shared school visionand develops one such vision. Chapters 4-6 offer a research-basedmodel of leadership as a problem-solving process. Chapters 7-9 aredevoted to selected aspects of problem solving in routine andnonroutine matters--the role of values, teacher development, andcreation of a collaborative school culture. The six chapterscomprising the third part address issues concerning the developmentof expert leadership. Chapter 10 assesses the value of a broad arrayof formal and informal experiences in developing expert leadership.Characteristics of formal preparation programs effective indeveloping expertise in routine and nonroutine matters are outlinedin chapters 11 and 12. The next two chapters examine howadministrator appraisal and selection processes might be designed andimplemented. Finally, chapter 15 proposes five broad strategies thatdistrict leaders might use in their efforts to foster the developmentof leaders for future schools. Six figures, 15 tables, and an indexare included. References accompany each chapter. The appendixcontains the principal profile-based instrument. (LMI)


U S DEPARTMENT OF EDUCATIONOIIK. of Eacanonai Researen ana improvamenn


This document has been roproduced as(eceived from the person or oroanizatmonuovinatmg it

MmOI changes nave been made lotrnprOve reproduction quality

PnInIS 01 OlIW Ot ODintOnS stated in Msur^e^t ^nt ren,ese,

'f,c,ai OE RI pos.bon or aohcv




The Falmer Press


Developing ExpertLeadership forFuture Schools


Developing ExpertLeadership forFuture Schools

Kenneth Leithwood, Paul T. Begley andJ. Bradley Cousins

The Falmer Press(A member of the Taylor & Francis Group)

London Washington, D.C.

UK The Falmer Press, 4 John St, London WC1N 2ETUSA The Falmer Press, Taylor & Francis Inc., 1900 Frost Road, Suite 101,

Bristol, PA 19007

© 1994 Kenneth 1.eithviood. Paul 1. Begle and J. Bradle, Cou,ins

.4/I rights reserved. No part of this publication may be reproduced, stored in aretrieval system, or transmitted, in any .form or by any means, elearonic,ituYhanind, photocopying, recording, or otherwise, without pernussion inu,riting .from the Publisher.

First published in this edition 1994

A catalogue record for this book is available from the BritishLibrary

Library of Congress Cataloging-in-Publication Data areavailable on request

1SWS 1 85000 743 8 easedISB\ 0 7507 0327 X paper

Jacket design by Caroline ArcherTypeset in 91/2/11pt Bemboby Graphicraft Typesetters Ltd, Hong Kong

Printed in Griut Britain by Bingos st ume Press, Basivstoke on paperwhich has a petaied pH value on final paper manufacture of not less than 7.5and is therefore 'aeid.free'.



List qf Figure..List of Tables viii

Acknowledgments ix

Foreword by Philip Ha llingerIntroduction

Part 1A Perspective on Developing Expert Leadership

for Future Schools

A Conception of Expert Leadership for Future SchoolsWhat Research Tells Us about thc Present State of SchoolLeadership



Part 2The Nature of Expert Leadership for Future Schools

3 Envisioning Future Schools 30

4 Expert School Leadership on the High Ground and in theSwamp 42

5 Expert School Leadership on the High Ground 55

6 Expert School Leadership in the Swamp 76

7 The Special Role of Values in School Leadership 98

8 Teacher Development: A Central Problem for Leaders ofFuture Schools 113

Collaborative School Cultures: A Key Part of the Solution 128

,Part 3Developing Expert Leadership for Future Schools

19 The Socialization of School-Leaders 148

11 Characteristics of Formal Programs for Developing Expert SchoolLeadership on the High Ground 166

12 Characteristics of Formal Programs for Developmg Expert SchoolLeadership in the Swamp 189


13 Performance Appraisal and Selection of School-LeadersPerformance Appraisal for Growth 205

14 Performance Appraisal and Selection of Cchool-Leaders: SelectionProcesses and Measurement Issues 221

15 Conclusion: Implications for District Leaders 246Appendix A 156Referetues 161Index 292


List of Figures

Fig. 2.1 An Orientation to Understanding Current School LeadirshipFig. 9.1 School-Improvement Processes: Causal Network for One

School 134

Fig. 13.1 Conceptual Framework of Evaluation Utilization 214

Fig. 14.1 Assessment-Center Dimensions and Descriptions 125

Fig. 14.2 Personnel Evaluation Standards by Knowledge Utilizationfactors 235

Fig. 14.3 Conception of Misutilizanon 237

List of Tables

Table 5.1 Expert School-leaders' Problem-Solving on the HighGround: A Summary of Levels 60

Table 5.2 School-leaders' Problem-Solving on the High Ground:Elementary vs. Secondary 71

Table 6.1 Expert School-Leaders' Problem-Solving Processes:Without Collaboration 84

Table 6.2 School-Leaders' Problem-Solving Processes: WithCollaboration 90

Table 7.1 A Classification of Values and Illustrative Statements 103Table 8.1 Interrelated Dimensions of Teacher Development 115Table 8.2 Development of Professional Expertise 116Tabk I0.1 Socialization Patterns Experienced by Aspiring Principals 149Table 10.2 Socialization Patterns Experienced by Practicing Principals 150Table 10.3 Patterns of School-Leaders' Socialization Experiences 155Table 10.4 Relationship Between Importance of Socialization

ctivities and Socialization Patterns 156Table 11.1 Objectives for Week One July 4-7 173Table 11.2 Relating Instructional Techniques to Learning Conditions 181Table 12.1 Development of Useful, Strategic Knowledge: Instructional

Strategies and Conditions 201Table 14.1 Instrument for Measuring School-Leader Expertise 230



Several of our colleagues have made quite dircct contributions to this book,although their names do not appear on thc cover. Doris Jantzi was responsible formuch of the data collection and analysis reflected in Chapter 9 and participatedextensively in the development of some of the ideas presented in Chapter 3.Rosanne Steinbach carried out much of the data collection and analysis reflectedin Chapters 6, 7, 10, and 12. Both Doris and Rosanne offered extremely helpfulsuggestions about the text as a whole.

Ail of thc word processing for the text was accomplished by Irene Switank-owsky. Her speed and accuracy contributed much to meeting our deadlines andkept editorial chores to a minimum.

Fini lly, significant portions of the research reported in the book were fundedby thc Social Sciences and Humanities Research Council of Canada. Fundingwas also received from the Ontario Ministry of Education through its blocktransfer grant to OISE. We are extremely grateful for their continued support.



In reading Developing Expert Leadership _for Future Schools I was stimulated toreflect on my experience of the past decade as a researcher and practitioner. As agra 'uatc student at Stanford University in the late 1970s and early 1980s I recallreading numerous reviews of research on school leadership. Surprisingly, thetone of these reviews contrasted markedly with the optimism engendered bystudies of effective schooling. While professional journals were replete withconfident. 'research-based recommendations for effective practice', respected re-searchers were considerably less sanguine about the depth and breadth of theknowledge base in school leadership.

For example, in 1982 Edwin Bridges published a critical review of researchin educational administration. He was unequivocably negative in his conclusions.

The state of the art is scarcely different from what seemed in place nearly15 years ago ... [Researchers! ... continue their excessive reliance onsurvey research designs, questionnaires of dubious reliability and valid-ity, and relatively simplistic types of statistical analyses. Moreover, theseresearchers persist in treating research problems in an ad hoc rather than aprogrammatic fashion. Equally disturbing is the nature of the know-ledge accumulated during this period. Despite the rather loose definitionof theory that was used in classifying the research ..., most !studies]proved to be atheoretical. Likewise, the research seemed to have littlepractical utility.'

For a practitioncr and fledgling researcher interested in administrator de-.velopment and school improvement, conclusions such as these were dishearten-ing. Researchers were not, however, alone in expressing th:ir dismay with thestate of leadership research and development. Several ycars later Roland Barthwrote an insightful critique, in which he lamented the misuse of research in thename of school improvement. His voice was that of a highly respected formerprincipal.

Our public schools have come to be dominated and driven by a concep-tion of educational improvement that might be called 'list logic' ... theintention of one state legislature is to 'identify competencies of effective

1 1


principals through research and develop training, certification, selection.and compensation procedures that recognize and support these com-petencies' ...

School improvement, then, is an attempt to identify what school peopleshould know and be able to do, and to devise ways to get them to knowand do it ... The list logic of educational change seems simple, straight-forward and compelling. Its only flaw is that it doesn't seem to workvery well.'

Barth refers here to the tendency of policymakers during the 1980s to reifyresearch findings and to employ them in ways unsupported by the research itself.Research results were disseminated in ways that ignored the craft knowledgeaccumulated by school-leaders and that failed to account for the varying contex-tual forces faced by leaders in real schools.

A decade later, Developing Expert Leadership for Future Schools seems to Justifythe cautious optimism of practitioners that principals can make a difference inschools while satisfying researchers' concerns about the validity of at least aportion of our knowledge base. The research, as well as the perspective, thatLeithwood, Begley and Cousins bring to bear on school leadership respond tolongstanding criticisms lodged by practitioners and researchers. As suggested bymy comments, this is no small feat. I will elaborate briefly.

First, unlike most research in our field, the studies on which this book isbased have a clear theoretical orientation. Conceptions of learning, informationprocessing, thinking and problem-solving developed by cognitive psychologistshave informed the direction of the authors' research and thc interpretation ofthcir findings. This theoretical orientation is interwoven in discussions of leader-ship and problem-solving throughout the book.

While the authors' research on expert leadership is theoretically informed, Itdoes not attempt to ape the social sciences for the sake of academic prestige. Thishook and its research base are explicitly grounded in 'problems of practice'. Thatis, while their selection of research problems has been influenced by cognitivetheory, both the form and substance of their work remains rooted in problemsfaced by live administrators in real schools. The book is primarily concernedwith what we now know about the problem-solving of school leaders and howthis might be applied to the improvement of schools.

The authors' respect for the craft knowledge of administrators is evident inthe nature of their research designs and in the role practicing administratorsplayed in generating and verifying their results. This is particularly apparent inChapters 6, 7 and 8 in which they discuss how administrators address 'swampy'problems, the ones our training programs seldom consider. Only direct interac-tion between researchers and practicing administrators could produce the resultsthey present. Readers will be heartened to know that the knowledge base for thisbook has benefitted from the numerous critiques and reality checks of practition-ers and other researchers.'

This leads to another notable feature of the book. It is based on 'programma-tic research' on the behavior and thinking of school-leaders. Ken Leithwood andhis colleagues were not the only scholars to review others' research on st hoolleadership during the early 1980s. However, to my knowledge. they arc the only


1. 2


researchers in the United States and Canada who persist in systematically follow-ing through with a program of original research that explicitly addresses thecriticisms identified in the published reviews. Their research program parallelsdevelopments in the field from the early 1980s, when the emphasis was onidentifying the behaviors of effective principals, to the early 1990s when the focusshifted to understanding the thinking that underlies those behaviors. Thus, thearguments presented by the authors about the nature of expert school leadershiparc based upon carefully developed propositions that have, for the most part,been tested in their own research. While Leithwood and his colleagues at OISEwould be the first to acknowledge the boundaries of this knowledge base, theeffects of their programmatic efforts to understand administrative problem-solving are undeniable.

One last facet of the book strikes me as especially worthy of mention: theauthors' grounding of expert leadership in an explicit vision of the future. Theirassumptions about the nature of leadership and the future context of schoolingare clearly stated for the reader to accept or reject. Their distinctions amongdifferent approaches to school leadership and the associated linkages to the needsof schools in the future seem compelling.

In conclusion, the ideas contained in this book have the potential to deepenthc reader's conception of school leadership, whether you are a veteran principal,an aspiring teacher or administrative leader, or a researcher. The concepts pre-sented by the authors will be particularly useful to those charged with assistingothers in developing the craft of school leadership. This belief is based upon myown experience. I have already begun to incorporate con :epts presented in thisvolume into my research, as well as into the training and development programsVanderbilt University offers to school-leaders. I have found the body of workrepresented in Developing Expert Leadership .for Future Schools useful and trust thatother readers will as well.

Philip HallingerSapphire Bay, St. Thomas



1 E. Bridges (1982) 'Research on the school administrator: The state of the art, 19671980', Educational Administration Quarterly. 18, 3, p. 25.R. Barth, (1986) 'On sheep and goats and school reform', Pin Delta Kappan, December,p. 294.

3 Readers will now that an article describing the program of research described in thisvolume was recognized as the pre-eminent piece of research published in thc EducationalAdministration Quarterly during 1990.

xiI 13


This is a book about change, change in the nature of schocl leadership as requiredby the nature of future schools. As with any change, the problem is one of gapreduction. In this case, the gap is between the nature of school leadership whichwill contribute productively to future schools and the nature of current schoolleadership. Accordingly, these are the issues addressed by the two chaptersincluded in the first of the three parts making up the book. Chapter 1 developsthe perspective on leadership for future schools explored in detail in Part 2 of thebook: this perspective is based on currently emerging ideas about expert lead-ership now and in the future. Chapter 2 describes what is known about thepresent state of school leadership, a description grounded in a systematic reviewof research on school leadership carried out over the past fifteen years. Takentogether, these two chapters clarify the extent of the change required in develop-ing expert leaders for future schools.

The second part of the book consists of seven chapters. They describe thenature of leadership we claim would be productive for future schools. A sharedvision of the future school, we suggest in Chapter 1, is a key to the potency offuture leaders. Chapter 3 raises several important issues involved in developingsuch a vision and, for purposes of illustration, develops onc such vision.

Chapters 4, 5 and 6 offer a research-based model of leadership as a problem-solving process: Chapter 4 clarifies the value and consequences of distinguishingbetween the solution of routine ('high ground') and non-routine ('swampy')problems; Chapter 5 describes effective school leadership on the high ground;and Chapter 6 describes expertise in the swamp.

Thc remaining three chapters in Part 2 are devoted to selected aspects ofproblem-solving on both the high ground and in the swamp. Chapter 7 examinesthe nature and role of values in the problem-solving of school-leaders. Contribut-ing to the growth of others in the school, as we argue in Chapter 1, is a majorcomponent of the leadership problem. Chapter 8 offers one conception of teacherdevelopment which may assist fulure school-leaders and outlines severalstrategies that seem promising in fostering such development. Creating a col-laborative school culture is one of the most powerful of these strategies: Chapter9 describes, in some detail how school-leaders might develop such a culture.

The six chapters comprising Part 3 address issues concerning the develop-ment of expert leadership. Chapter 10 identifies a broad array of formal and


informal experiences potentially contributing to such development and estimatestheir relative value. Characteristics of formal preparation programs effective indeveloping expertise on the high ground and in the swamp are outlined inChapters 11 and 12 respectively. Chapters 13 and 14, using a commonframework, examine how administrator ppraisal and selection processes might bedesigned and implemented in order to contribute more than is typical to thedevelopment of expert leadership. Finally, Chapter 15 proposes five broadstrategies which district leaders might use in their efforts to foster the develop-ment of leaders for future schools.

We have tried to take up the issues addressed in this book in a serious butreadable way. The book is a product of the past four years of a research programcarried out at OISE's Centre for Leadership Development. With few exceptions,each chapter is based on our own original research within that program and/or acareful review of the research of others. In each case, research methods are brieflydescribed but no effort is made to provide the amount of detail and justificationone would expect in an academic journal. Readers of the book, we assume, areinterested primarily in our results, after some assurance that those results are theproduct of careful inquiry. Additional sources are noted for those with moreextensive, methodological interests.

Throughout the book we use, more or less interchangeably, the termsprincipal, vice-principal and school-leader. Conceptually, we agree that leader-ship may be exercised by virtually anyone in the school, although hard evidencesuggests serious limitations and obstacles for those not in formal leadershiproles. However, our treatment of school leadership is research-based and mostsuch research (ours and others) has been carried out with principals and vice-principals. In reality, then, most of our claims about school leadership are claimsabout leadership exercised by those in principal or vice-principal roles. Further-more, throughout the book the term 'school-leader' is applied to those whomight aspire to the actual exercise of leadership, whether or not they realize thataspiration.



A Perspective on Developing ExpertLeadership for Future Schools

i C

Chapter 1

A Conception of Expert Leadership forFuture Schools

Of all thc hazy and confounding areas in social psychology, leadershiptheory undoubtedly contends for top nomination. Probably more hasbeen written and less is known about leadership than any other topic inthe behaviorai sciena:es. (Bennis, 1959, p. 259)

Adding a future's twist to the leadership story, as we do in this book, wouldhardly make Bennis more confident about what we write. But we are fearless.We possess the kind of self-confidence that Klemp and McClelland (1986) labeleda 'generic' leadership skill which seems fitting (the Greeks called it 'hubris'but no matter).

Schools of the past, present and future needed, and will continue to need,competent management. They need people who can establish and maintain thedaily routines that make individual people in the organization dispensable thatallow the basic purposes of thc school to be achieved even though members ofthe school inevitably change. Schools also need to change. And for change toresult in improvement, schools require expert leadership. This book is primarilyabout leadership, but it is also abc.ut how to carry out managerial work in a waythat contributes to leadership.

This chapter is intended to clarify the purposes of the book. It is also anopportunity to share with you some of the assumptions and perspectives wehold, insofar as they have shaped our thinking about future schools and leader-ship. Five assertions provide a framework for doing this:

schools are durable institutionsschools are instruments of social changeschool-leaders are key artisansthe leadership problein has three partsthe leadership process is usefully viewed as problem-solving

Schools are Durable Institutions

Potential readers, we suspect, will quickly judge us to be either hopelesslytrapped in a mind-set shaped by current institutions or conservatively optimistic.

4 17


A Conception of Expert Leadership for Future Schools

And the judgment may well be rendered even while reading the title of thisbook. If you believe that schools are brittle, bureaucratic anachronisms incapableof meaningful adaptation to the brave new world of the twenty-first century, thisbook is not for you. After all, we use the term 'future schools' in our title,whereas for you, the term (like 'airplane) food' or 'pleasant nightmare') is anoxymoron. Give your old Ivan Illich (Mich, 1971) another read instead.

Conservatively optimistic is our own self-concept. We believe schools areimperfect but enduring and improvable institutions that society will not beprepared to do without, for longer than any of us are likely to be around. Afterall, schools in western societies have borne the brunt of a more or less bad press,in cycles, throughout at least this century. In the face of a constant barrage ofboth constructive criticism and hopelessly uninformed carping, many schoolshave responded with continuous, adaptive changes while still preserving theiressential form and function. That is a formula for successful evolution whetherthe organism be biological or social. Our general purpose in this book is to makea modest and indirect contribution toward the subsequent evolution of schoolsby influencing the development of those who will be their leaders.

Schools are Instruments of Social Change

Concerns about what the future holds for schools are everywhere; they may befound, for example, in the successive 'waves of reform' initiated by states in theUnited States (e.g., Bacharach, 1988), in the dramatic efforts to refocus the locusof control over schools in Australia (e.g., Marsh, 1988) and the United Kingdom(e.g., Walford, 1990), and in the comprehensive curricular reforms now beingintroduced in some parts of Canada (e.g., Province of brit:sh Columbia, 1989).The strength of these concerns and their undeniable impact on schools is largelyexplained by a renewed belief in the link between education and the achievementof fundamentally important social goals for the future. Some view this link aseconomic salvation through educational excellence. As Drucker (1989) asserts: 'ina world whcre knowledge has becomne the true capital and the premier wealth-producing resource, the process of education is the ultimate supplier of power'.While rejecting the values of perpetual growth and materialism on which theeconomyeducation link is forged, others also look to education as a key to theirpreferred image of the future. One such image, developed by Ornstein andEhrlich (1989), concerns itself with the long-term, collective well-being of thespecies, a well-being dependent upon appreciating the fragile nature of the planetand upon a set of values and lifestyles in harmony with its fragile nature. And, inOrnstein and Ehrlich's view, with such an image in mind, 'refashioning howpeople are educated could have enormous import for thc future of the species'(cited in Mitchell, 1990, p. 29).

School-Leaders are Key Artisans

If education, in general, and schools, in particular, are seen as tools for socialchange, educational leaders are assumed to be among the most critical artisans.This assumption is widely held by the public-at-large, as well as by education


1- 0

Developing Expert Leadership for Future Schools

professionals (e.g., Schlecty, 1990). It is also an assumption warranted by re-levant evidence (e.g., Immegart, 1988). Indeed the 'leadership effect' becomesincreasingly prominent the more one focuses attention on schools as opposed toother types of organizations. Research on effective schools (e.g., Wilson andCorcoran, 1988) is especially clear on this point. Developing school-leaders,therefore, is one of the most promising avenues available for successfully address-ing the changes which will challenge future schools. A more specific purpose ofthis book is providing insights from research about the development of leaderscapable of facilitating the changes which will be required of schools in the future.

The Leadership Problem Has Three Parts

There are many people who are able to get out in front of the band whohave nothing to do with what song the band is playing. (Corbett,Wilson and Aducci, 1990, p. 1)

Having a vision of what they would like their schools to be in the future iscritical for school-leaders; it may even put them in front of the band. But it is(among other things) the creation of a shared vision among those playing theinstruments that determines what song is being played, and whether it is one ormany. With this as a critical task, it is reasonable to ask whether the front of theband is the best place for the leader to be. Our conception of the leadershiprequired for future schools suggests that the rear of the band and the midst of theband will offer opportunities that are at least equally as important as opportun-ities available at the front.

George Terry (1960) defined the leadership problem as 'how to influencepeople to strive willingly for group goals', which pretty much sums up themessage contained in our band metaphor. Two components of the leadershipproblem are evident in this definition: one focuses on how to influence people(the process of leadership); the other focuses on determining the goals towardwhich influence is exercised (the intended product of leadership). Leaders need tobe able to assist individuals, often working in groups, to identify agreed-upongoals, the framework for which ought to be a vision of a future school. Leadersalso need to be able to influence the same individuals and groups to strivewillingly for the achievement of these goals. This is what Sergiovanni (1987,p. 121) refers to as 'leadership by purpose'. As he points out:

Purposing is a powerful force bccause of our needs for some sense ofwhat is important and some signal of what is of value.... The objectof purposing is the stirring of human consciousness, the enhancement ofmeaning, the spelling out of key cultural strands that provide bothexcitement and significance to our work.

Exercising influence, the process component of leadership, involves theexercise of power. It is useful, for present purposes, to distinguish among foursources of power. Two of these include: (a) the power that comes from theauthority vested in the position (jurisdictional powers) held by the leader (e.g.,principal, superintendent), and (b) the power from support provided by those



A Concepuon of Expert Leadership for Future Schools

who are pursuing their own interests (e.g., political power). Leaders relying onthese sources of power usually spend most of their time at the front of the band.However, these two sources of power are less and less available, at least forschool-leaders now and in the future. Not many principals, for example, attri-bute much vested authority to their position for a variety of reasons: for example,the growing strength of teacher unions, and the trend toward decentralizingeducational decisions (school-based management, school-based curriculum de-velopment), which seems to have as its primary goal the empowerment ofteachers, not school-leaders. These obvious, pragmatic reasorls for school-leadersto reduce their reliance on traditional sources of power ar.. consistent in theireffect with efforts to build teacher leadership into a more powerful force in futureschools (Smylie and Brownlee-Conyers, 1990). These reasons are also consistentwith efforts to reconceptualize leadership from a feminist perspective. Reflectingon her own intellectual voyage as a school-leader, Regan (1990, p. 568) forexample, speaks of the need to give much greater attention to those aspects ofleadership which involve 'caring and nurturing relationship, and communitybuilding' or the 'soft' side of leadership. Similarly, studies of district leadershipare beginning to suggest that:

a strong superintendent in future years is less likely to be a 'take charge'boss than an 'unheroic' and more consultative leader ... working withothers ... facilitating, finding common ground, listening and persuad-ing. (Crowson and Morris, 1990, p. 41)

These are what Crowson and Morris refer to as 'considerative qualities of cultureand choice'.

Such views of leadership in thc future will be based on two other sources ofpower for influencing people. One of these is the power that is awarded by virtueof one's content or technical expertise (e.g., knowledge about schooling and skillin performing valued functions). To be used effectively, this source of power isbest exercised in the midst of the band, in concert with thosc considerativequalities alluded to above. The purpose for exercising such power is to assistdirectly members of the school community to overcome the obstacles they facein striving for their vision of the school. In Leiberman and Miller's (1990) terms,this is what it means to be a leader of leaders, a person whose power of expertiseis used to achieve ends rather than control people. Management and leadershiparc intertwined in the midst of thc band. Given a vision of what it means to dothe right things, content or technical expertise focuses on doing those thingsright.

A final source of power comes through the ability to empower others,something often best accomplished from the rear of the band. Leadership basedon this source of power is often referred to as 'transformational' (e.g., Burns,1978, Bass, 1985, Sergiovanni, 1990) or developmental (Schlecty, 1990). Theterm 'transform' implies major changes in the form, nature, function and/orpotential of somc phenomenon; applied to leadership, it specifies general ends tobe pursued although it is largely mute with respect to means. From this begin-ning, we consider the central purpose of transformational leadership to be thcenhancement of individual and collective problem-solving capacities of organiza-tional members; such capacities are exercised in the identification of goals to be


Developing Expert Leadership for Future Schools

achieved and practices to be used in their achievement. As Bennis and Nanus(1985, p. 217) clarify, leaders are transformative when they arc able to `shape andelevate the motives and goals of followers'. Such leadership:

is collective, there is a symbolic relationship between leaders and follow-ers and what makes it collective is the subtle interplay between thefollowers' needs and wants and the leader's capacity to understand ...these collective aspirations.

Transformational leadership, thus, is culture changing (Coleman and LaRocque,1989). Specific ways in which school-leaders effectively solve this part of theleadership problem are explored in Chapter 9.

Technical expertise exercised in a 'considerative' context and the ability toempower others are the two sources of power through which future school-leaders are most likely to influence others, and which they must be prepared touse. These are also the most defensible sources of power for leaders of organiza-tions, dedicated to such fundamental values as respect for individual persons,representative democracy, and professionalism with its emphasis on clientwelfare.

The leadership problem, in sum, has three parts: developing a widely shared,defensible vision; in the short run, directly assisting members of the schoolcommunity to overcome obstacles they encounter in striving for the vision; and,in the long run, increasing the capacity of members of the school community toovercome subsequent obstacles more successfully and with greater ease. Schoolsoperate in a dynamic environment which exerts constant, often contradictorypressures for change: future schools are likely to experience even greater press-ures of this sort. For this reason, future school-leaders will have to respond tothese problems in what Vail (1989) refers to as 'permanent white water'. Turbu-lence will be the norm not the exception.

The Leadership Process is Usefully Viewed as Problem-Solving

A recent review of studies of leadership and leader behavior in education, byImmegart (1988), concluded that research on educational leadership has actuallydeclined over the past decade. This has been the case in spite of the increasedchallenges which have confronted schools during the 1980s and will continue andlikely escalate throughout the 1990s. Furthermore, much greater attention needsto be given to the conceptualization of educational leadership, according toImmegart (1988), attention that acknowledges the multidimensional nature ofleadership. Our response to this current state is to offer an initial conception ofleadership as problem-solving, and to elaborate on the dimensions of that con-ception throughout the chapters of Part 2 of this book.

Problem-solving is a productive conception of leadership for two reasons.First, problem-solving is a generic human function, and as a result, capable ofhelping unearth the roots of otherwise puzzling human activity. As a concepionof leadership, problem-solving competes with a host of more superficial concep-dons as identified, for example, by Bass (1981): leadership as a focus of group



A Conception of Expert Leadership for Future Schools

processe-, personality and its effects, a form of persuasion, a power relationship,the exercise of influence, and the initiation of structure (and some five others).

Problem-solving is also a productive conception of leadership because itrepresents a plausible next step along the path that has been trod by those activelyengaged in leadership research and theory. According to Immegart (1988), thispath began with efforts to uncover personal traits of leaders (e.g., intelligence,dominance, self-confidence, high energy level). Subsequent steps down the pathfocused on leader styles, and then, through efforts to operationalize styles, onspecific leader behaviors (a step on which studies of leadership, emanating fromthe effective schools' research, stalled). Current conceptions include efforts tounderstand the situational or contingent nature of leadership: that is, the extent towhich a particular leader act and styles or behaviors depend for their effects onthe context in which leadership is to be exercised. The central skill of situationalleaders is to decide to choose, from their repertoires, responses that are called forby the circumstances: `Given what I know of this teacher's ability and disposi-tion, and given the placement request being made by this parent, I am going toturn down the parent's request without further consultation with anyone'.Decision-making, however, is only one form of problem-solving and among thesimplest forms, at that. Faced with relatively simple and routine problems, whatschool-leaders do is captured reasonably well by the notion of decision-making.But a more comprehensive understandirw of the full range of intellectual andemotional activity which constitutes problem-solving is required to appreciatethe responses of leaders to complex and non-routine problems.

In education, the dominant conception of school leadership with which ourproblem-solving conception competes at present, is `instructional leadership'.This term symbolizes the importance, to school leadership, of an emphasis onstudent growth, and on much of the direct service provided by schools infostering student growth. Such an emphasis was wholly appropriate and timelyto bring to school leadership during the early 1980s, when the term gained awidespread following. But 'instructional leadership' conveys a meaning whichencompasses only a portion of those activities now associated with effectiveschool leadership. This is quite apparent, even in texts which explicitly adopt'instructional leadership' as their thematic orientation: for example, Duke's (1987)School Leadership and Instructional Improvement, and Smith and Andrew's (1989)Instructional Leadership. These texts explore many issues of importance to schoolleadership (including visions of effective schools, for example) that fall outsidethe bounds of `instruction' as it is normally conceived. Instructional leadership, aswell, seems not to adequately encompass what has now been learned about theneed for school-leaders to redesign professional work cultures to support teachergrowth (e.g., Roscnholtz, 1989), as a context for instructional improvement, forexample. Nor does this term (instructional leadership) acknowledge thc contribu-tion of non-instructional features of schools (e.g., informal, non-instructionalrelations between teacher and individual student) to their attractiveness to stu-dents typically considered at risk of dropping out (Rumberger, 1987, Lawton andLeithwood, 1988). Furthermore, schools of the future may well provide a signi-ficant portion of their services through means not readily associated with instruc-tion as presently conceived (an image of the teacher as resource, rather thaninstructor, might be appropriate for future schools).



Developing Expert Leadership for Future Schools

Summary and Conclusion

We began this chapter by affirming our assumptions about the durability ofschools as future institutions, about their instrumental role in social change, andabout the significant contribution of leadership to the functioning of schools,now and in the future. We then argued that the problem for leadership in thefuture had three parts: developing a shared, defensible vision of a future schoolconsidered desirable by those with a stake in it; directly assisting members ofthe school in addressing the challenges encountered in their efforts to achieve thevision; and increasing the capacity of school members to address those and futurechallenges themselves, more successfully. Problem-solving, we suggested, is anappropriately generic and comprehensive conception of what will be demandedof future leaders, more appropriate, for example, than 'instructional leadership'.

Our view of leadership, as problem-solving, encompasses dispositions andorientations toward leadership, now being given special attention in the leader-ship literature. It encompasses, for example, both transactional' and transfor-mational leadership processes; it reinforces the continued importance of thecontent expertise of school-leaders, usefully highlighted in the term 'instructionalleadership'. But it extends the scope of the school-leader's role to thinking like anorganizational designer. Such thinking expands the array of problems givenexplicit attention by school-leaders to include the underlying structural andcultural conditions of work in the school and how they influence the nature ofservice provided to students.


l The approximate translation of 'hubris' is overweening self-confidence.2 Transactional leadership is based on exchange theory. It assumes that the leader moti-

vates followers by exchanging various kinds of incentives, for the cooperation offollowers in working toward organizational goals.

10 23

Chapter 2

What Research Tells Us about thePresent State of School Leadership

Assuming the principalship of a school becoming the formal school-leaderappears to involve taking on certain requisite duties and challenges. Everythingwe know about the role suggests that it is hectic and fast-paced, involvingsignificant amounts of interpersonal contact, of which more is unplanned thanplanned. And this seems to be the case for virtually all school principals. Indeed,at the close of many days, most principals would express considerable supportfor the applicability of 'chaos theory' (Gleick, 1987) as an explanation for theirwork: the metaphorical butterflies, flapping their wings in the Far East, seem tohave created unique and unpredictable 'weather patterns' in Joan Fitzgerald'sschool in Newfoundland. But this is as close as we come to identifying naturallaws of school leadership.

Beyond what might be called this ecology of the role (and we don't wish tominimize its importance), everything else is up for grabs. Which is to say thatformal school leadership is a socially constructed role, the expectations for whichhave changed dramatically since its inception. Recently, expectations havechanged at a sufficiently rapid rate to create incompetence among some of thosewith long tenure in the role. That is, at some earlier point in their careers, theperformance of these people matched the socially determined expectations forexemplary school leadership. But the social ground shifted from under them andthey did not shift with it. When planned change is defined as a process ofreducing the gap between current and desired states, sometimes you have to runhard to stay in the same place. This happens when the desired state changes fasterthan you do. Under such circumstances, if you only amble forward, you actuallylose ground!

'What is the purpose', you ask, 'in acknowledging the problem of changeand the socially constructed nature of expectations for school leadership?' De-veloping leaders for future schools, in our view, ought to be considered aproblem of planned change at two levels: at the level of how school leadership forfuture schools is conceptualized (e.g., what qualities are associated with suchleadeship) and at the level of how individual people can be assisted in acquiringthose capacities or qualities needed to exercise leadership in future schools. Whilemost of this book addresses change at the individual level, such change dependson being clear about change at the conceptual level. At both levels, clarity aboutthe changes which need to be made depcnds on defining the ends of the planned



Developing Expert Leadership for Future Schools

Figure 2 1- An orientation to understanding current school leadership

2. School-Leaders'1 Influences Practices

3 Impact

2.2--> Classroom

2.1 Factors


1 1 1 2Actions

2.3 Effects on>External Mental - School-wide Students

Influences Processes Factors and Staff

change gap. Chapter 1 provided an overview of what we conceptualize to bedesirable leadership for future schools. This chapter' clarifies what can be learnedabout the current state of school leadership from existing research. This researchhas focused almost exclusively on principals and vice-principals, and so, theseroles are the exclusive focus of this chapter.

Understanding Current School Leadership

An adequate understanding of current school leadership depends on much morethan what is provided, for example, by descriptions of overt leadership be-haviors. Figure 2.1 identifies one important set of variables and relationships thatare central to such an understanding: it concerns the nature, causes and consequ-ences of what current school-leaders do. An earlier version of this frameworkwas proposed as a consequence of Leithwood and Montgomery's (1986) reviewof research. The framework is also similar to one proposed by Dwyer et al.(1984). The Figure is intended to suggest that what school-leaders do - their'practices' (component 2) - is most directly a consequence of what they think -their mental processes (component 1.2). Such mental processes are a function ofcertain characteristic ways of understanding, applied to the environment in whichthey work. Elements in this environment (component 1.1) may be interpreted byschool-leaders in many ways and certain elements turn out to have much greaterimpact on their thinking than other elements. School-leaders have been observedto engage in quite distinct patterns of practice (component 2.1), shaped by howthey think about their work. Three or four such patterns are particularly evidentamong North American principals and these patterns have demonstrated widelyvarying impacts on classroom practices or factors (component 2.2) and factorswithin the school but outside the classroom (component 2.3). It is primarilythrough such influence that school-leaders affect pupils and others (component3), although some aspects of their practices are more direct than others in theirimpact.

As suggested by the orientation to current school leadership represented inFigure 2.1, this chapter addresses three sets of questions: questions about theimpact, practices, and influences on the practices of current school-leaders. These

12 25

[Aar Research Tells Us about the Present State of School Leadership

sets of questions are addressed through a relatively comprehensive review ofempirical research reported between 1974 and 1988. The review was conducted intwo parts. For research reported between 1974 and 1984, three literature reviews(examining a total of seventy-five original empirical studies) (Leithwood andMontgomery, 1982, Leithwood, 1982, Leithwood and Montgomery, 1986) wererelied on. For the 1985 to 1988 period, sixty original studies were identifiedthrough systematic search techniques described in detail in Leithwood, Begleyand Cousins (1990).

We turn now to a report of what the results of these studies of elementaryand secondary school principals (135 in total) contribute to the three questions ofconcern in this chapter. I.Z.!sults of the more recent studies (1985-1988) aredescribed in more detail tnan results of the 1974-1984 studies. Consistent withour practice in all chapters of the book, we make quite explicit the evidence used,in order to address each question. In addition, we also identify directions forsubsequent research which would expand the knowledge-base relevant to the taskof developing leaders for future schools.

What is Known About the Impact of Current School-Leaders?

The three reviews of pre-1985 research identified forty empirical studies ofelementary principals' impact on some aspect of students and teachers. Withrespect to students, the impact included: positive attitudes toward school (twostudies), achievement in basic reading and math skills (fifteen studies), andreduced vandalism and absenteeism (two studies). Principals impacted onteachers' job satisfaction (seven studies), use of innovative practices in the class-room (eleven studies), and teachers' perception of the principal's leadership (fivestudies). Seven studies conducted between 1985 and 1988 provided additionalevidence concerning principals' impact on students' basic skills, teachers' jobsatisfaction, and their use of innovative teaching practices. Four additional typesof impact on teachers were explored in these studies.

Andrews el al. (1986) used the gain scores of students on standardizedachievement tests to explore the effects of principals rated by their teachers asstrong, average, or weak leaders. Significant correlations were found betweenachievement and strength of leadership for both math and reading gain scores. Inthis study, strength of leadership was a function of (a) the extent to which theprincipal mobilized personnel and other types of resources to achieve the school'sgoals, (b) the clarity of communication concerning the school's goals, (c) theextent of active involvement in the school's instructional program, and (d) the ex-tent to which the principal was a visible presence in different parts of the school(teachers were given these criteria).

Blase, Dedrick and Strathe (1986) used teachers' responses to questions aboutthcir principal's behavior, stress caused by their principal, and the principal'simpact on their classroom performance, to explore correlations with teacher jobsatisfaction. A moderately strong association was found between teacher satisfac-tion and the degree to which principals' initiation of structure (the extent towhich a leader initiates, organizes and defines work to be done and the manner inwhich it will be done) and consideration behaviors (behaviors related to enhanc-ing teachers' self-esteem) were perceived to help teacher performance. Brady

2 6 13

Developing Expert Leadership for Future Schools

(1985) found 'principal supprtiveness' or consideration, in particular, to be themost consistently significant predictor of staff perceptions concerning thc preva-lence of group, as opposed to individual, decision-making in the school. Suchsupportiveness, in the teachers' view, was also related to: principals' involvementin curriculum decisions, the use of an interactive (rather than an objectives-driven) curriculum-planning model, intimacy (e.g., good social and interpersonalrelations) among staff, and satisfaction with the school curriculum.

Sharman (1987) explored the relationship between the degree of teachers'implementation of a new math program, and principals' evaluation, supervisionand staff development initiatives. Results suggested that the more directly suchinitiatives were seen to support implementation, the greater the level of use of theinnovation by teachers.

Loyalty, the extent to which teachers are committed to the principal andhave an unquestioning faith and trust in the principal, was the nature of theimpact of interest in Johnson and Venable's (1986) study. Different types ofprincipals"rule administration behavior' (e.g., democratic vs. authoritarian) andprincipals' influence in the school-system hierarchy were types of school-leaderpractices related to such impact. Results suggested differences among elementaryand secondary teachers in their reaction to different types of rule administration.Greater loyalty among elementary teachers was most closely related to less'7,mishment-centered' rule administration (less conflict, less tension, and lessexplicit enforcement of rules) by principals. More representative rule administra-tion (joint rule, initiation, and acceptance) was most closely related to secondaryteachers' loyalty to the principal. The loyalty of both groups of teachers wasassociated with their perception of the principals' ability and willingness to exertinfluence upwards in the school-system hierarchy and to do things for theicachcr.

Hoy and Brown (1986), like Blase et al. (1986) and Brady (1985) examinedthe effects of principals' consideration (e.g., attention to interpersonal relations)and initiation of structure (e.g., attention to the task and how to get it done).Both aspects of leadership were found to be related to the teachers"zone ofacceptance' (their readiness to accept decisions made for them by the principal).Together, these two sets of behaviors accounted for 38 percent of the variance inteachers' zone of acceptance. As with Johnson and Venable (1986), differencesbetween elementary and secondary teachers were found: secondary teachers attri-buted overriding importance to the principals' initiation of structure.

What do these studies have to contribute, in sum, to the question of currentschool-leaders' impact on schools? And what new research would be helpful tosupport leaders of future schools? First, we must acknowledge significant limita-tions in the research-based knowledge about the nature of current school-leaders'impact. But based on the number of studies alone, one can reasonably concludethat current school-leaders are capable of having a significant influence on thebasic skills' achievement of students. A recent review of school effects in theThird World also attributes such impact to school-leaders (Fuller, 1987). As well,current principals seem capable of influencing teachers' adoption and usc ofinnovative classroom practices, and teachers' job satisfaction. Evidence concern-ing other types of impact is extremely thin, however.

Several suggestions for subsequent research which would support leaders of

14 27

Wliat Research Tells Us about the Present State of School Leadership

future schools are worth mentioning. First, research to date has been concernedwith the principals' impact on an important but highly restricted set of studentoutcomes (attitudes toward school, basic skills, vandalism and absenteeism).Such outcomes reflect neither the scope nor the emphasis of the full range ofstudent outcomes to which many schools aspire now and, certainly not, in thefuture (see Chapter 3 for more on this). Such schools, for example, now oftenview basic math and reading skills as instrumental in fostering growth in higherorder thinking skills and in the acquisition of complex, discipline-based conceptsand theories. Many schools, now and in the future, assume responsibility forassisting students in the development of social and attitudinal outcomes (e.g.,self-concept, esteem for the culture and customs of others) of special importance,in light of changing family and community contexts. The research-knowledgebase required to assist leaders of future schools will need to inquire into how suchleaders can impact on outcomes of this sort, outcomes which better reflect thelikely mission of future schools.

A second suggestion for research concerns the nature of principals' impacton teachers. There is no underlying, comprehensive theory dictating the choiceby researchers of what types of impact on teachers to examine. Further, with thenotable exception of teachers' use of innovative practices, all of the research todate has focused on attitudes and dispositions only loosely linked to teachers'performances. Of more value to leaders of future schools would be a choice ofteacher outcomes, driven by a theory or theories of teacher growth in classroomeffectiveness, such as the one described in Chapter 8. As we argue in that chapter,contributions to teacher development are central aspects of the job of leaders forfuture schools.

Third, perhaps the school characteristic currently of most interest in effortsto understand effective schools is school culture, or ethos (the norms, values,beliefs, and associated behaviors shared by those in the school). This characteris-tic is not independent of students and teachers. It is, however, a more compositefeature of schools that cannot be understood by looking only and separately atstudents and teachers. While schools which vary in effectiveness also appear tovary in the nature of their culture, it is not clear whether principals can signi-ficantly influence school culture. This makes culture a promising focus for atten-tion in subsequent research on principals' impact. Chapter 9 explores, in moredetail, some quite recent research which gives warrant for optimism about therole of school-leaders in respect to school culture. This chapter is relativelyspecific about what school-leaders should also do to foster the development ofmore productive school cultures.

What is the Nature of Current School-Leader Practices andHow Do They Vary?

Seventy-five pre-1985 studies provided information about principals' practices.Of thc thirty 1985-8 studies with such a focus, four inquired into roles, fivedescribed principals' overall patterns or styles, seven focused on the practices of'typical principals'. and fourteen studies examined the practices of highly effectiveprincipals.


Developing Expert Leadership for Future Schools


Prior to 1985, research on the principalship included efforts to clarify principals'roles, beginning from two quite different premises. One premise was that therole could be viewed as predominantly unidimensional and the research objectivewas to discover the dimension which best captured the role. Principals, forexample, were claimed to play a largely 'manager' role or a largely 'leadership'role: they were concerned mostly with administration or with instructionalleadership. Results of this research usually found typical practice consumed bymanagerial or administrative tasks, but desired practice best captured in leader-ship roles focused on such substantive educational decisions in the school asinstruction.

Other research on principals' roles, however, was based on the premise thatthe role was multidimensional. Sergiovanni's (1984) five `leadership forces' illus-trate reasonably well the range of dimensions the principal's role was found toencompass in this research. These dimensions included: technical managementactivities, provision of interpersonal support and encouragement to staff, instruc-tional intervention, modeling important goals and behaviors, signaling to otherswhat is important (symbolic leadership), and developing an appropriate andunique school culture.

All four of the 1985-8 studies of principals' roles assumed a multidimension-al view of the role. Brubaker and Simon (1987) inquired about the actual andpreferred roles of principals from among five possibilities: principal teacher,general manager, professional and scientific manager, admthistrator and instruc-tional leader, and curriculum leader. Each of these roles was dcscribed in para-graph length and included at least several dimensions of practice. Most principalsviewed their current roles as administrator and their preferred role as instruction-al leader. General manager was rated a distant second, as current role, andcurriculum leader and professional scientific manager tied for second choice, aspreferred role. Gender differences emerged in this study, with women givingmuch higher ratings for actual and preferred roles to administrator, instructionalleader and curriculum leader. Men, in contrast, rated the general manager rolemuch higher than women, as both an actual and preferred role.

Gousha's (1986) survey also found highest ratings given to instructionalleadership as a description of actual and preferred roles. Other roles rated highlywere school manager, personnel leader and disciplinarian. These role ratings werenot consistent with principals' estimates of the time spent on five key tasksassociated with their role. School management and teacher/student concerns(personnel leadership) consumed 40.8 percent and 34.1 percent respectively.School improvement (instructional leadership), on the other hand, consumedonly 15 percent of their tune.

The disciplinarian role was the special focus of attention in Montgomerie,McIntosh and Mattson's (1987) study. Opinions were solicited from teachers,superintendents, principals, and board chairs, concerning the relative importanceof roles played by principals. The framework for this study was a modifiedversion of Sergiovanni's leadership forces: thc disciplinary role was added, andcultural and symbolic forces were collapsed Results of this study, combining theopinions of the four groups of respondents, gave strongest weight to the symbo-lic, disciplinarian and humanistic roles, and lea st weight to the instructional and



What Research Tells Us about the Present State of School Leadership

technical roles. Teachers, however, showed a strong preference for the disciplina-rian role of the principal.

A fourth study of roles by Bradeson (1986) identified three metaphors forthe role adopted by principals and others with whom they interacted. The roleappeared to be dominated by a metaphor of maintenance: principal, as the personwho sees and understands the total process and is responsible for keeping theprocess going. About three-quarters of the time of principals was devoted tomaintenance tasks, and about 5 or 6 percent of this time was devoted to tasksassociated with each of two additional role metaphors: survival and vision.Survival tasks were those focused on meeting such immediate needs as short-range planning. The ability of the principal to holistically view the present, toreinterpret to all its constituents the school's mission, and to speculate aboutfuture directions was Bradeson's meaning of vision.

Patterns or Styles

Research aimed at describing patterns or styles of principal practice has examinedsuch practice in more depth than the roles' perspective has: it has attempted eitherto identify dominant orientations to the role, without concern for differences inimpact, or to define progressively more effective styles or patterns of practice.Results of four pre-1985 studies using this approach can be summed up in fourleadership styles which we refer to as A, B, C and D.

Leadership style A is characterized by a focus on interpersonal relationships,on establishing a cooperative and congenial 'climate' in the school, and effective,collaborative relationships with various community and central office groups.Principals adopting this style seem to believe that such relationships are critical totheir overall success and provide a necessary springboard for more task-orientedactivities in their schools.

Student achievement and well-being is the central focus of leadership styleB. Descriptions of this class of practices suggest that while such achievement andwell-being is the goal, principals use a variety of means to accomplish it. Theseinclude many of the interpersonal, administrative, and managerial behaviors thazprovide the central focus of other styles.

Compared with styles A and B, there is less consistency, across the fourstudies reviewed, in the practices classified as style C (program focus). Principalsadopting this style, nevertheless, share a concern for ensuring effective programs,improving the overall competence of their staff, and developing procedures forcarrying out tasks central to program success. Compared with style A, theoricntation is to the task, and developing good interpersonal relations is viewedas a means to better task achievement. Compared with style 13, there is a greatertendency to view as a goal the adoption and implementation of apparentlyeffective procedures for improving student outcomes, rather than the studentoutcomes themselves.

Leadership style D is characterized by almost exclusive attention to what isoften labeled 'adininistrivia': the nuts and bolts of daily school organization andmaintenance. Principals adopting this style, according to all four studies, arepreoccupied with budgets, timetables, personnel administration, and requests forinformation from others. They appear to have little time for instructional and



Developing Expert Leadership for Future Schools

curriculum decision-making in their schools and tend to become involved only inresponse to a crisis or a request.

Hall et al. (1986) argue that their three styles (responder, manager, initiator)have different effects on the process of school improvement. Initiators are moresuccessful in their school-improvement efforts, responders are least successful.Paragraph-length descriptions are provided for each of these styles.

In order to better understand the specific practices associated with each of theHall et al. styles, Stevens and Marsh (1987) inquired about principals' vision andstrategies for achieving their vision. Results suggested that more effective styleswere associated with better integrated visions more directly focused on program-related matters and with a greater number of them. More effective styles alsowere associated with a greater range of strategies and more effort in theirstrategies to focus on a combination of daily, small-scale, and comprehensivelarge-scale changes.

Research by Leithwood and Montgomery resulted in a much more detailed(chapter length) description of four multidimensional patterns of practice,ordered from least to most effective in accomplishing a complex array of studentoutcomes. The patterns arc labeled: administrator (least effective), humanitarian,program manager, and systematic problem solver (most effective). In Chapter 4,these patterns are described in much greater detail, as are increasingly effectiveapproaches to solving predictable, routine school-leadership problems.

Three additional studies since 1985 have focused on principals' styles moreor less directly based on a conception of leadership, provided by the Ohio StateLeadership Studies in the 1960s. Consideration and initiation of structure are thetwo dimensions defining this conception. Hoy and Brown's (1986) survey sug-gested that high degrees of both principals' consideration and (especially) initia-tion of structure influenced teachers' readiness to accept decisions made for themby principals (their 'zone of acceptance'). Blase et al. (1986) reported similarresults in relation to teacher satisfaction and classroom performance. Teachersand principals responding to Brady's (1985) survey attributed substantial import-ance to the supportiveness (or consideration) of principals in fostering a variety ofdesirable attitudes among staff.

Typical Practice

Judged by the quantity of research available, interest in describing multiplepatterns or styles of practice has been quite restricted. In contrast, there has beena relatively large number of studies designed to describe and understand both'typical' and 'highly effective' forms of practice. Of the seventy-five empiricalstudies conducted between 1974 and 1984, fifty-two (69 percent) were concernedwholly, or in part, with the nature of typical practice.

Studies of typical practice usually differ from studies of principal roles andpatterns of practice in terms of the detail of information they provide. Leithwoodand Montgomery's 'dimensions' of principals' practices or behavior arc a usefultool for bringing some conceptual coherence to such detailed descriptions. Thesedimensions include (a) goals principals attempt to achieve in their schools (nature,source, and use of such goals), (b) factors in classrooms and the school which


What Research Tells Us about the Present State of School Leadership

principals believe they must influence to accomplish their goals (choice of factors,nature of expectations held for factors, source of these expectations), (c) strategiesused to influence such factors (criteria for choosing strategies, emphasis amongstrategies, characteristics of strategies), and (d) the nature of decision-makingprocesses.

Descriptions of 'typical' practice, while sometimes quite detailed, are consis-tent with the pattern of practice described above as leadership style D, with someelements of A. All but one of the seven recent studies provided informationabout at least one of the four dimensions: goals, factors, strategies, and decision-making. Kingdon (1985) compared expectations for the role of the full-timeteaching principal, on the part of such principals, with expectations normallyheld for the role; other aspects of their activity were also examined. Few differ-ences in expectations were found but teaching principals did give first priority totheir teaching assignments and did most of their administrative work outsideregular school hours.

Bradeson's (1986) was the only study to provide information about thc goalsof the typical principal. Such information was available in his analysis of thepurposes served by carefully recorded daily communications. These purposesand the percentage of communications devoted to them by principals were(a) maintenance messages, concerned with policies and procedures (49.8 percent),(b) human messages, concerning peoples' attitudes, morale and satisfaction (25.6percent), (c) task messages, concerned with the quality and quantity of education-al services (23.7 percent), and (d) innovation messages, concerned with schoolimprovement (1 percent) These four sets of purposes correspond closely to thefocuses of major concern in each of the four leadership styles that have beendescribed. 'Running a smooth ship', both organizationally and interpersonally,appeared to preoccupy the typical principal.

The single study (Chater, 1985) of principals' decision-making, among theseven reviewed, was conducted in seven secondary schools in Great Britain. Noeffort was made to identify relative levels of effectiveness of the principals(heads). Results suggested that while most principals sought relatively low in-volvement from staff on financial matters there was some variation amongschools in other decision areas. Many staff were satisfied to have low levels ofinvolvement in school decisions because it reduced their uncertainty. Only oneschool used highly participatory forms of decision-making.

Three studies provided information about principals' strategies. Focusing oncommunication patterns, Bradeson (1986) found considerable variation in the

location of principals' communication. Almost three-quarters of principals' activ-ities were interpersonal and took place with only one other person, over halfinvolving face-to-face contact. The main thrust of these results was replicated byDavies (1987) in Great Britain and by Gaily (1986) in Israel. In fact, typicalprincipals' activities in most countries, where data are available (e.g., US, Cana-da, Australia, Great Britain, Israel), appear to be characterized by brevity, frag-mentation and variety.

Four studies touching on factors principals' influence provided minimalinformation about practices within this component. Williams (1986) inquired intoprincipals' influence on fourteen components (factors) of teachers' instruction.Classroom composition, teaching materials and resources, and instructionalmethods were factors principals perceived themselves to influence. Instructional


Developing Expert Leadership for Future Schoolc

methods, including student grouping, facilities, and the community were do-mains (factors) used by Gaily (1986) to explore the location of administrativeactivities. The extent to which principals influenced these factors was not ex-plored in this study. A similarly designed study (Taylor, 1986) in Great Britain,although not directly investigating specific factors influenced by principals, sug-gested that heads spent nearly half their time dealing with classroom factors. Thisis not typical practice in most countries and may be a function of the traditional'head-teacher' role of the school-leader in Great Britain. Ehiametalor (1985)examined the actual and expected levels of performance of principals in Nigeria,in influencing factors classified as curriculum and instruction, staff and students,the community, school organization and structure, and budget. Principals wereclassified by age, experience, and training. For the most part, all categories ofprincipals performed below expected levels; performance was substantially high-er, however, for principals with twelve to nineteen years experience.

Effective Practice

A relatively large amount of research activity between 1974 and 1984 (fifty-onestudies) was devoted to studies of effective practice. Results of this researchprovide a detailed account of what has been described above as leadership style B(a student-achievement focus) with elements of style C (program focus). Thirteenstudies of effective practice conducted between 1985 and 1988 were includedin this analysis. These of these studies explicitly compared typical and effectivepractice. By far the largest proportion of attention (twelve studies) was devoted,in these studies, to strategies used by principals: four studies spoke to the goals ofeffective principals, two compared factors, and three described aspects ofdecision-making. With respect to goals, highly effective principals were found todemonstrate high levels of commitment to goals for the schools, especiallyinstructional goals. They ensure thar school instructional goals are congruentwith district policies. Such principals articulate an overall vision for the schoolwhich is multifaceted. This vision emerges from a belief that all children canlearn what the school has to offer. Effective principals set relatively high profes-sional and school standards for goal achievement and actively work toward thedevelopment of widespread agreement concerning such standards (Taylor, 1986,Dwyer et al., 1984, Andrews et al., 1986, Larsen, 1987).

Two studies provided information about effective principals' approaches toclassroom and school factors. Larsen's (1987) data suggested substantial efforts toinfluence the classroom curriculum, teachers' instructional behaviors, materialresources for instruction, and the general environment of the school (climate orculture). Dwyer's (1984) study identified seven such targets for principals: workstructure, staff relations, student relations, the environment, plant and equip-ment, community relations, and institutional relations.

The relatively large number of studies identifying strategies used by effectiveprincipals generated twenty-two such strategies. Ten of these strategies wereidentified in just one study and are not reported here. Those strategies identifiedin three or more studies included: monitoring student progress; teacher evalua-tion and supervision; establishing and communicating clear, high expectations forstudents and staff; establishing and enforcing an equitable discipline code; and



What Research Tells Us about the Present State of School Leadership

maintaining a positive school clirnate. Strategies associated with effective princip-als in two studies included: goal setting; planning and program development;mobilizing and allocating resources; modeling; being actively involved in staffdevelopment for teachers and self; and developing good working relationshipswith staff, community, and central office staff.

Two of the three studies touching on effective decision-making processes ofprincipals provide additional support for already well-established claims concern-ing the benefits of participatory decision-making. Stanard's (1986) case study of asingle principal attributed a portion of the principal's success in solving disciplineproblems to her involvement of parents and staff in both curriculum and disci-pline decisions. Johnson and Venable (1986) found that participatory forms ofdecision-making (representative rule administration') were associated with grea-ter teacher loyalty to the principal among at least secondary teachers; the datawere less conclusive with respect to elementary teachers.

High and Achilles's (1986) data were partly at odds with the general supportfound for participatory forms of decision-making. Their study inquired intoprincipals' and teachers' preferences, for the use of seven different bases of social

power by the principal. Teachers ranked highest bases of power labeled: expertpower, legitimate authority, and norm-setting power. Principals awarded morepotential to involvement and less to legitimate authority bases than did teachers.These results prompted High and Achilles to comment in their conclusions:

Principals in general have apparently been reading too much of the 1960sliterature (togetherness) and believing it. (1986, p. 15)

Given thc image of desirable, future school leadership, loosely framed inChapter 1, results in this section appear to offer useful insights for developingfuture school-leaders. There are holes in these results which, if filled, would alsobe of value.

The significance of research to date, on roles, appears to be threefold. First,the school-leader's role, now and in the future, is clearly multidimensional andfurther efforts to identify the 'most important' dimension would be misdirected(Pitner and Hovecar, 1987). Second, among the many role categories used asframeworks for research, Scrgiovanni's (1984) five 'leadership forces' sup-plemented by a disciplinary category, seem to represent available data as well asany. Filially, in light of the more detailed knowledge about principals' practicesgenerated from other perspectives, a role perspective no longer offers a usefulframework for subsequent research aimed at helping develop future school-

leaders.Available research on patterns or styles of practice supports the claim that

school-leaders carry out the job in distinctly different ways. Most of thesedifferences are well represented by four focuses: a student achievement focus, aprogram focus, an interpersonal focus, and a focus on routine maintcnanceactivities. Furthermore, these focuses appear to constitute levels of effectivenessin which the main concerns defining lower levels (e.g., a focus on routinemaintenance) are incorporated into, and subsumed by, the conccrns defininghigher levels (e.g., a student achievement focus). Additional empirical tests of thcclaim that the four patterns of practice represent a hierarchy of effectiveness arcneeded, as is a more detailed description of how school-leaders come to adopt


Developing Expert Leadership for Future Schools

certain patterns of practice. Siich description is basic to the task of developingfuture school-leaders and the results of our own, just completed research on thismatter are outlined in Chapter 10.

The extensively researched dimensions of leadership labeled considerationand initiating structure are important dimensions within each of these four levelsor styles, consideration being the main concern of principals with an interperson-al focus. But further research within the limitations of theoretical and methodo-logical frameworks traditionally used to explore these dimensions cannot bejustified: the importance of the two dimensions is no long:- in question andyou will recall the special emphasis we attributed to consideration in Chapter 1.Detailed knowledge of practice within each dimension is what is still lacking andwould be of value for developing future school-leaders.

Recent studies of typical practices reinforce but do not extend prior know-ledge about such practices. Such studies paint a surprisingly uniform picture ofsuch practive across many national contexts. Heads in Great Britain were some-what unique in their orientation to classroom factors. With empowerment andschool improvement as goals, more research simply describing typical practicedoes not seem likely to be of much use in the development of future leaders.

Results of research on effective current practice appear to bc quite useful indescribing in more detail the qualities valued in future school-leaders. Thisresearch confirms the central role that principals' goals play in und.:rstanding thesource of effective practice. These goals form a central part of the vision princip-als use to bring consistency to an otherwise unmanageably diverse set of de-mands. Developing a widely shared, defensible vision is central to future schoolleadership, as we discussed in Chapter 1. Effective principals act to influence abroad array of school factors with an extensive repertoire of strategies. This isusing technical expertise in the midst of the band, as also discussed in Chapter 1.Their priorities are expressed in their day-to-day actions; they are better attuned,than are typical principals, to behaviors that actually influence teachers. Effectiveprincipals use participatory decision-making, selectively but frequently, depend-ing on their assessment of the context. Nevertheless, as Pfeifer (1986) has noted,much of the data on effective practice has been generated in the context ofturbulent, urban schools. Further studies of effective practice in diverse contextsare essential if results are to be used with confidence as guides to practice, in abroad array of settings, in the future.

What Influences the Practices of Current School-Leaders?

Results of research concerning influcnces on current principals' practices arcuseful in beginning to think about what sorts of experiences will contribute mostusefully to future school-leader development (or growth). Influences on currentprincipals' practices were examined using two sets of research. f he first setincluded eighteen empirical studies previously analyzed by Leithwood and Mont-gomery in their 1982 literature review: results of these studies are only brieflysummarized in this section. The second set, receiving more attention, includedstudies reported from 1982 through 1987, with greatest emphasis on the 1985 to1987 period. These studies were organized according to their conceptualization ofindependent, mediating, and dependent variables, as follows:



What Research Tells Us about the Present State of School Leadership

Set A: Studies which examine the relationship between external influences

and principals' practices (Previous reviews of literature spoke only to this set

of relationships)Set B: Studies which examined the relationship between external influences

and principals' internal mental states or processes

Set C: Studies which inquired about relationships between internal mental

states or processes, and principals' practicesSet D: Studies which examined relationships among all of external in-fluences, internal mental states or processes and principals' practices.

Set A: External Illuences and Principals Practkes

Leithwood and Montgomery's review of eighteen relevant studies identified five

classes of 'obstacles' standing in the way of principals providing instructional

leadership (the dependent variable); four of these were external to the principal.

Obstacles presented to principals by teachers included: lack of knowledge

and skill about new practices; uneven professional training; and lack of mo-

tivation to change, to participate in in-service training, and to collaborate in

planning. Obstacles also identified were teacher autonomy and constraints on

program decision-making resulting from collective bargaining and union contracts.

Several features of the principals' role were viewed as obstacles: ambiguity

(unclear expectations, conflict about responsibilities) and complexity (number of

people to consider, number of tasks). Hierarchical structures and problems they

created in making changes were characteristics of school systems, identified as

obstacles to principals. So too were excessively rigid and time consuming policies

and procedures; provision of inadequate resources; and conservative stance of

central administrators toward school-initiated change. We address influences such

as these and their consequences for district leadership in more detail in Chapter

15. Finally, aspects of the community were also viewed as a source of obstacles to

principals in their efforts to be more effective. These included: the interests of

parents (too much or too little), pressure of special interest groups in the com-

munity, and excessively conservative views about the nature of appropriate

school programs.Four 1985-8 studies provided support for the general thrust of the results

reviewed by Leithwood and Montgomery. In a follow-up study by Leithwood

and Montgomery (1984), principals reported having only moderate concerns

about the four sets of obstacles (above) as a whole. Ob5tacles associated with

school districts appeared to present the greatest difficulties, a finding also re-

ported by Goldman and Kempner (1990). but no strong relationships were found

between classes of obstacles and principals' effcaiveness. Obstacles associated

with the district also dominated evidence presented by Crowson and Morris, and

Louis. Crowson and Morris (1985) suggested that in one large urban school

district, between a half and a third of principals' time was consumed in respond-

ing to formal, hierarchical controls largely having to do with budget, personnel,

and pupil behavior. Informal reward systems provided by the system (e.g.,

getting a better school, promotion) attracted considerable additional nine

of principals. Louis (1989) also reported a strong but indirect influence by

super-intendents and other district office staff on the planning and design of

3 C

2 3

Developing Expert Leadership for Future Schools

school-improvement efforts in a large sample of US secondary schools. While thedistrict was not the dominant source of problems in school-improvement efforts,conflicts with district office staff, staff turnovers, competing priorities for change,and eroded school autonomy were viewed by a large proportion of principals asserious challenges to their school-improvement initiatives. The main thrust ofthese findings was supported by Gousha (1986).

Leithwood and Stager's (1986) study of problem-solving processes suggestedthat highly effective principals, with administrative experience, become morereflective about their own processes and refine these processes with time.Although similar to moderately effective principals in general moral values and inpersonal values, effective principals are more influenced by their beliefs concern-ing principals' roles and responsibilities, and are more able to specify day-to-day consequences of such beliefs; they arc also more aware of school-systemneeds and requirements, and try harder to take them into account in school-level problem-solving. Effective principals derive more personal enjoyment fromproblem-solving, and partly as a consequence of this, are more proactive indealing with school problems. Chapter 6 expands on, and adds to, these resultsin reference to future school-leaders.

Finally, Tracy (1985), and Brubaker and Simon (1987) linked differences inthe socialization experiences of men and women with differences in career aspira-tions and view of the principal's role. Such experiences appear to cause more mento seek the principalship earlier in their careers (before the age of 30) and to aspireto the superintendency as a career move. Gender-related socialization experiencesalso seemed to contribute to a relatively large proportion of women viewingthemselves, as already noted, more as curriculum and instructional leaders. Re-latively larger proportions of men, in contrast, viewed themselves as generalmanagers. Greater amounts of formal education were also associated with atendency for principals to view themselves as curriculum and instructional lead-ers. These matters are explored in detail in Chapter 10 as they bear on developingfuture school-leaders.

Set B: External Influences and Internal Mental States or Processes

Eight studies (1985-8) provided data on the relationship between external in-fluences and internal mental processes. Two of the eight studies concernedperceived job stress or feelings of 'burnout' (Sarros and Friesen, 1987, Kotkampand Travlos, 1986). Volume of work, poor interpersonal relations with staff andothers, pressures from higher authorities and role conflict were external factorsappearing to contribute to feelings of burnout.

Three studics examined a variety of external influences on principals' jobsatisfaction (Caldwell and Paul, 1984, Sparkes, 1986, Gunn and Holdaway, 1986).External influences identified in these studies, contributing to positive attitudestoward the job, included larger schools and communities, length of experiencein the role, and more qualification/training. Such influences also included highlevels of teacher ability, cooperative teacher attitudes, recognition by others ofone's work, and relatively lower levels of conflict and workload.



Set C: Internal Mental States or Processes and Principals' Practices

Five studies conducted since 1985 which focus on this area of the principalshipwere located. The independent variables in thcse studies were principals' beliefs,values, and problem-solving processes.

Taylor (1986) reported a strong association between the effectiveness ofprincipals, and their belief that all students can learn. Principals' use of student-achievement data in decision-making was associated, by Glasman, with three setsof beliefs by principals concerning their control over the use of such data, and itsvalue in program and teacher evaluation.

Using a hierarchy of values proposed by Hodgkinson (1978), Begley (1988)inquired into the role of such values in principals' decisions to adopt micro-computer technology in their schools (discussed more fully in Chapter 6). Prin-cipals with greater knowledge of the innovation and those with an instructionalorientation to their roles were more likely to make their adoption decison usingthe value 'consequences for students'. Other principals more often based theirdecisions on their personal preferences, a desire for consensus (e.g., among staff),

or some broad moral principle. Values of consequence increascd as the basis for

choice with increased knowledge among all types of principals. Subsequentresearch suggests a particularly crucial role for values in the work of futureschool-leaders. As a consequence, Chapter 7 is devoted to a more detailedexamination of that role.

Leithwood and Stager compared problem-solving processes used by 'highlyeffective' principals with those uscd by more 'typical' principals. Results ofthese studies, we believe, are central to our expectations for future school-leaders. Hence, further discussion is saved for the detailed treatment contained in

Chapter 6.

Set D: External Influences, Internal Mental States or Processes and

Principals' Practices

The five studies in this set address three problems. The first probLmi, addressedby Darcsh (1987), and Marshall and Greenfield (1987) concerns effectiveness inthe early years of the principalship. These studies suggest that reduced effective-ness during these years as a principal and one's unwillingness to take risks is a

direct function of inadequate skills (internal states) in: (a) carrying out routineadministrative procedures, (b) conflict management, and (c) determining system-

wide decision-making processes. It is also a function of feelings of dissonance

with one's values and reduced excitement about school improvement. Suchfeelings and skill deficiencies are, in turn, attributed to external influences, such

as: restricted administrative experiences as a vice-principal, inadequate formal

training for the role, and socialization processes prior to assuming the principal-

ship.Effective principals differ from their less effective peers, in part, in terms of

the extent and quality of information used in their decision-making. McColskey,Altschuld and Lawton (1985) inquired into the reasons for variation amongprincipals in this component of their practice. "Fraining in social-science research



Developing Expert Leadership for Future Schools

methods appeared to be an important external influence on such practice. Open-mindedness and beliefs concerning the principal's role and the autonomy andpower available to effect change in the school were internal influences identified.

Cousins (1988) studied principals' use for professional development ofappraisal data concerning their own performance. He found that principals'attitudes toward the appraisal processes werc predictive of the extent to whichthey learned about their performance. Attitudes were found to be associated withhigh levels of motivation for professional growth and inversely related to prin-cipals' experience and working knowledge. Use of appraisal data for decision-making was found to be linked to external variables (e.g., nature of decision tobe made, communicative aspects of the appraisal process). In Chapter 13, weextend the insights available in this and other research to an exploration ofselection and appraisal procedures appropriate for leaders of future schools.

Finally, in the context of implementing policies initiated outside theirschools, Trider and Leithwood (1988), and Leithwood (1986) found significantdifferences in influences on principals' practices depending on: (a) principals'orientation to the role, (b) stage in the implementation process, and (c) principals'training and/or policy-relevant knowledge. More instructionally oriented prin-cipals were less influenced by district factors and more guided by their ownbeliefs. As implementation proceeded, organizational context factors (e.g.. staffinput to decisions) within the school took on greater significance for all princip-als, to the extent that such factors had the potential for solving emerging prob-lems of implementation in the school; so, too, did the support available fromvarious groups outside the school. This was consistent with evidence reported byFullan, Anderson and Newton (1986) in relation to secondary-school principals'efforts to implement a major instructional innovation in their school. Finally,school administrators with specialized knowledge in the policy drea being im-plemented made decisions in a relatively autonomous fashion, guided largely bytheir own beliefs. These findings suggest the possibility, in fact, that principals'special knowledge (often the result of training) is one of the central determinantsof the pattern of policy-implementation behavior in which they engage. Princip-als without special knowledge seemed to rely extensively on the guidance pro-vided by central office staff and the existing skills of staff. Perhaps their concernfor working relationships can be attributed to their dependence on knowledgepossessed by others, and their desire to gain cooperation from such people toapply that knowledge in their school.

By way of summary, current school-leaders' practices are influenced by fourtypes of external factors: the principals' role (e.g., expectations, complexity), alarge cluster of influences concerning the attitudes, abilities and behaviors ofothers (e.g., teachers' willingness to innovate), characteristics of the school sys-tem (e.g., district policies and procedures), and the principals' own 'background'(e.g., training, socialization experiences). These external factors interact withprincipals' internal mental processes and states: personal traits (e.g., openminded-ness). knowledge and beliefs (e.g., about what is best for students), values (e.g.,consequences for students). attitudes and feelings (e.g., job satisfaction), andskills (e.g., problem-solving, conflict management). Through such interactionsthe specific nature and effectiveness of principals' practices are shaped.

As we reviewed these influences above, we pointed to results of specialconsequeme for future school-leader development and identified subsequent

26 39

What Research Teils Us about the Present State of School Leadership

chapters in the book where such matters are taken up in depth. There are someobvious holes in the research results, however, with the development of futureschool-leaders as a goal. First and most generally, this body of research evidencecurrently available is extremely limited in quantity. It is also uneven in qualityboth conceptually and methodologically, for instance, there are no experimentalstudies exploring cause-and-effect relationships, and little of the research is

guided by a coherent theory to explain or suggest relationships among thevariables of interest, if choices must be made. There is, arguably, a greater needfor research exploring these relationships than there is more descriptive researchon effective practice, for example. Such research would help us understand how

effective practice develops, a crucial matter about which current research has little

to say. What factors lead a school-leader to adopt a particular one of the four

dominant orientations to the role (A, B, C, 1))? Can school-leader's dominantorientation change? If so, how does this happen? How can the experiences offuture school-leaders be constructed so as to contribute best to the developmentof effective orientations to the role? Productive answers to these questions, wesuspect, are rooted in a better understanding of principals' internal mental pro-cesses and states: the rational aspects of these processes, such as the content andorganization of knowledge structures, as well as such non-rational elements asbeliefs, attitudes and values.

Further research, addressing issues discussed here, also seems likely to clarity

some of the fundamental reasons underlying differences in practices observedamong male and female principals. There is some urgency about developing abetter understanding of this matter, since much larger numbers of women arenow actively preparing themselves for leadership in future schools.


The purpose of this chapter was to sum up what is presently known fromresearch about current school leadership. Comparing the results of this researchwith the conception of leadership for future schools, described in Chapter 1,provides a clearer sense of the nature and size of the change required to movefrom current leadership practices to those suitable for future schools. Such acomparison leads us to several broad conclusions. First, typical current school-leadership practices are woefully inadequate, given present expectations for

schools and leaders never mind the considerably more ambitious expectationsanticipated for schools of the future. Redesigning the experiences that producesuch patterns of practice is a crucial developmental strategy. Such redesigninvolves not only individuals' s cialization experiences directly (see Chapter 10)

but also the wider organizational structures which account, to a significant

degree, for those experiences (see Chapter 15). Second, practices currentlyviewed as effective have much to offer to leadrs of future schools. But thesources of such practices are not well understood. Because these sources unavoid-ably involve what has been termed, in this chapter, 'internal processes'. efforts todevelop future school-leaders will need to focus much more directly on suchprocesses. Chapters 4-7 describe what is presently known about these processes.Chapters 11 and 12 describe formal programs for developing such processes.


Developing Expert Leadership for Future Schools

Finally, knowledge about school leadership at present reveals little or no-thing about those transformational aspects of the role, identified in Chapter 1, asso important for leaders of future schools. While this may be due to inadequaciesin the research base, there is little doubt, as well, that transformational leadershipis poorly understood and rarely practiced. Developing the individual and organ-izational capacity for exercising transformational leadership is one of the mostsignificant challenges in developing leaders for future schools.


1 This chapter is based on a comprehensive review of literature by Leithwood, Begleyand Cousins (199(l)


Part 2

The Nature of Expert Leadership forFuture Schools

Chapter 3

Envisioning Future Schools

The source of the authority of the practical intelligence ... is derivedfrom some kind of overall vision of society. Every person with anyfunction in society at all will have some kind of ideal vision of thatsociety in the light of which he operates. One can hardly imagine asocial worker going out to do case work without thinking of her ashaving, somewhere in her mind, a vision of a better, cleaner, healthier,more emotionally balanced city, as a kind of mental model inspiring thework she does. (Frye, 1988, p. 70)

While Northrop Frye is undoubtedly correct, not all mental images of an idealtomorrow are created equal. What this means is nicely illustrated in Steven's(1986) study of the visions held by elementary principals who differed in theirbasic patterns or styles of practice, and their resulting effectiveness as described inChapter 2. This study found that principals with more effective patterns or stylesof practice had more extensive, detailed, and integrated visions than did princip-als with less effective styles. Furthermore, those with the most effective style(e.g., student-oriented, referred to as Style B in Chapter 2) had the strongestprogram-related visions and worked hardest and most persistently to changetheir schools in the direction of their visions. It is reasonable to infer from suchresults that the nature of principals' visions, as well as their efforts to implementthose visions, help account for their impact on schools.

The Virtues of Vision

Predicting the future is a hazardous business, as our daily surprises about oilcrises, political events in eastern Europe, holes in the ozone layer, and the like,attest. But so long as we aspire to some control over our own destiny, we haveto continuously give it our best shot. Some aspects of our lives, in any event, canbe shaped significantly by our aspirations, and the nature of schools is amongthem, to a point. No matter the type of organization (e.g., Morgan. 1989, Vail,1989) or the conception of leadership (e.g.. Bass, 1981, Chapter 1), a vision orpicture of what the organization ought to be seems to be vital to success,


4 3

Envisioning Future Schools

especially during turbulent times. As one of the executives in Morgan's studynoted:

The world is such a changeable place that you need to have a wellarticulated long-term sense of where you're going, which gives you thebase, the confidence to take on whatever adaptability issues come alongwithout losing your sense of direction. You've got to respond to theissues of the moment without losing that long-term sense. (1989, p. 46)

Useful visions provide relatively precise guides to action and allow reason-ably subtle discrepancies in need of attention to be detected in one's school. How,for example, can school-leaders interpret those aspects of a situation encounteredin the schools which deserve their attention without a standard for comparison?The nature and size of the discrepancy between one's vision of 'what ought tobe' and one's perception of 'what is' constitutes problem interpretation. With-out a defensible standard for comparison one has to rely on others to tell themwhat issues are worth attention. Educational leaders use their vision of thehealthy school much as a physician uses an understanding of the healthy,well-functioning body. With the vision or silhouette of excellence in mind, theschool-leader assesses the organization. A defensible vision makes for responsibleautonomy in a leader. It is also the basis for proactivity, for determining priori-ties about how to spend time now, for setting clear goals, and for other aspects ofplanning. Useful, defensible visions are the product of careful thought, systema-tic effort and continuous evaluation and refinement. They are not the fluffyproducts of armchair daydreaming which the term itself suggests and manycurrent administrators seem to believe.

To be of greatest value in informing the work of school-leaders and theircolleagues, a vision must be widely shared. Such sharing will usually be theproduct of much collective thought and discussion. That vision will also reflectunique aspects of the school and thc community which it serves. This is not tosay, however, that a vision is entirely context-bound or informed only bysources within the school's environment. As one among several other sources,research has an important role to play in identifying elements of a vision whichwarrant adoption across many schools. Results of research can also providestimulating points of departure as school-leaders begin to work with their com-munities and professional colleagues to. build a sense of direction. Broad butpredictable social and professional trends are other sources useful in informingone's vision.

Developing a useful, defensible, shared vision in these terms, while criticalto do, is not a simple business. It is a do-able business, nevertheless, althoughyou may be doubtful about this point. Anticipating such doubt, the remainder ofthis section illustrates what we mean by broad social and professional trends andhow they may he used in shaping a vision. We also demonstrate the uses thatmight be made of relevant research and what we mean by detail and precision inreference to a leader's vision. The product of our effort is one vision of futureschools in which the horizon is about ten ycars from now. Other visions arepossible, of course. Duke (1987) for example offers an alternative with aspects%...hich arc both similar to. and different from, what is described in this chapter.


Developing Expert Leadership for Future Schools

Most of the differences are traceable to the use of different information, notsurprisingly.

Illustrative Social Trends and How They Might Be Reflected in a Vision ofFuture Schools

Among the many social trends which may influence the nature of schools duringthe 1990s, five seem especially powerful and their effccts predictable. One ofthese is a trend for the economies of most developed nations to continue to moverapidly from an agricultural and industrial base toward an information technolo-gy base (Hay and Roberts, 1989). This trend seems likely to increase the valueattached by the public to education since specialized skills, especially amenable toschooling, are required in such economies. It also seems likely to create a need forschools to give much more serious efforts than in the past, for accomplishinga more ambitious and complex sct of goals (e.g., higher order thinking skills)along with more flexible and efficient curricula and instruction for their achieve-ment. The availability of information technology provides, as well, opportunitiesfor greater organizational diversity in schools: for example, a division of instruc-tional labour between teachers and computers; more site-based decision-making,made possible through ready access to information previously accessible only tothose in the central office.

A second social trend of consequence for schools is a gradually aging popula-tion, a population with little vested interest in the education of youth but withquite direct interests in life-long learning opportunities for themselves. Schoolswishing to respond to such interests will need to address a broader range ofgoals, a quite different set of goals than those appropriate for pre-adult clients.Also required will be more flexible, client-centered forms of instruction, and awillingness to collaborate with other education agencies in providing meaMngful

services.Increasing cultural, religious, and ethnic diversity within communities is a

third, well established trcnd likely to continue through the 1990s and to haveprofound effects on schools. Such diversity creates a more ambiguous context inwhich to decide on school goals and priorities. A broadened range of intellectualand non-intellectual goals will have to be considered by the school in response toa broader range of values and aspirations in the community. Changes in forms ofeducation also seem warranted not only by such broadening of goals but also as aresponse to cultural norms which do not condone some of the behaviors associ-ated with 'tried-and-true' forms of instruction in western society (e.g., publicdisplays of knowledge, competition among individuals). Ethnic diversity alsoplaces pressure on personnel-selection criteria: the need for the ethnic composi-tion of the teaching force to reflect that diversity, at least in some measure.

Increasing respect for the rights of individuals independent of income, class,ability, religion, age. race or sex is a fourth trend of consequence for futureschools. This trend is being given considerable force not only through the courtshut also through the efforts of many advocacy groups, and even through thechanges to the constitution of some nations, Canada among them. This trendreinforces the legitimacy of equity as an educational goal, especially regardingequal access to knowledge. not just educational resources. Perhaps the main force

.12 4'

Envisioning Future Schools

of this trend is, and will continue to be, felt on instructional practices: teacherswill need to master a larger, more flexible repertoire of strategies, and suchpractices as ability grouping, will come under irresistible criticism as discrimina-tory, however attached education professionals might be to such practices fortheir own reasons. Greater respect for individual rights is changing and willcontinue to change the ethnic and gender balance in positions of responsibility inschools; many more members of minority groups and women will assume suchpositions, shortly.

Finally, the gradual but dramatic shift underway for the past twenty years inthe nature of families and their role in the education of youth will continue tocreate sometimes subtle but nevertheless quite compelling demands on schools.For example, James Coleman's (1983) concept of 'social capital' helps point outthat what schools do best provide students with intellectually demanding tasks,participation in which makes growth possible depends on a stable, supportivefamily environment. Such an environment gives students the self-esteem andself-confidence to respond to such challenges, as they arc intended. Yet, there hasbeen a clear trend, likely to continue for some time, toward the destabilization oftraditional family forms and the increase in proportion of new family structures.Many of these new structures do not provide the social capital for children thatschools assume, as they attempt to carry out their role. A successful response byfuture schools to this change will include even greater attention to the socio-emotional needs of students and thc provision of nurturance in a form notuncommon in many elementary schools, at present, but quite rare in mostsecondary schools.

Many more trends than the five discussed here will influence schools overthe next ten years. But from just this small selection of trends, it is plausible toenvision future schools, in sum: addressing a broader, more complex set of goals;using forms of instruction not much in evidence at present and possessing alarger repertoire of such strategies; and possibly serving a group of clients otherthan their traditional youthful clients. It also seems plausible to envision futureschools with the technical resources to make a larger portion of their owndecisions, in collaboration with other education agencies and with a professionalstaff more closely matching the gender and ethnic composition of the studentpopulation.

Illustrative Professional Trends and How They Might Be Reflectedin the Vision

Schools, of course, have been immersed in efforts to change over the past decadein response to aspects of the social trends discussed above, as well as others.These efforts have kicked into play a number of movements, within thc profes-sion, which have sufficient momentum to be viewed as independent sources ofinfluence on schools throughout this decade: school-leaders will need to at leastconsider what part, if any, these movements will play in the vision developed fortheir schools. McDonnell (198)) provides a useful, fourfold classification of thesemovements as they have been experienced in the United States. They includegreater decentralization of authority and empowerment of teachers (a nmvementto.4ard site-based inanageincntIlso strongly advo(ated in the UK and Australia)


Developim Expert Leadership for Future Schools

and increased public accountability (experienced almost everywhere as evidencedby increased attention given to student testing and the establishment of 'bench-marks'). Additionally, there are substantial efforts being made to alter thecontent and process of classroom instruction and to strengthen the links betweenschools and their communities.

Unlike broad social trends, the implications for schools of these professionaltrends are obvious and quite direct. Depending on one's schocl setting, some ofthese trends already may have the weight of policy. For example, versions ofsite-based management arc legislated in the UK and Australia, and a potentiallyempowering form of shared decision-making (staff collegial councils) is a re-quireinent for all schools in the Canadian province of British Columbia. Formost schools, however, significant portions of these professional trends remainoptional, and school-leaders and their colleagues must consider whether or not toinclude them in their vision of the future school.

Incorporating Research Results into the Vision

Most research describes sonic aspect of the present or the past. One might wellask: 'How is that useful in formulating a vision for the future?' The mostpractical answer we have to this question (there arc others: see Morgan, 1989, forexample) depends on the acceptance of an intermediate term definition of thefuture, such as the ten-year horizon we have adopted. With such an horizon, it isreasonable to claim that research describing especially effective current practice isuseful. By definition, such practices are found rarely and in those cases, usuallyonly piecemeal. Implementing such practices, more or less comprehensively, inour future school would be an ambitious task to undertake in a tcn-year timeframe, especially as those practices arc modified to reflect the broad social andprofessional trcnds referred to earlier. This line of reasoning assumes a gradualevolution of aspirations for schools (not the imposition of a radically new set) andan incremental approach to educational change.

In order to illustrate what the outcome would be of using a relevant body ofresearch in formulating a vision of future schools, we reviewed twenty originalstudies of effective secondary schools. These studies were supplemented byseveral research reviews of effective elementary schools in order to identifypossible differences between effective secondary and elementary schools. Thetwenty original studies were identified through a process of ERIC searches andbibliographic follow-up, and on prior knowledge of the authors, selected byusing predetermined criteria.

In addition to school leadership, six categories of characteristics or dime-nsions within which secondary schools appear to vary in effectiveness were evid-ent in the studies reviewed. They include the goals given priority by the school,the attributes and practices of teachers and school-leaders (omitted from discus-sion in this section, but discussed in subsequent chapters), the nature of schoolprograms and classroom instruction, school policies and organizational features,school culture, and the nature of school-community relations. These categories ofcharacteristics appear to have either direct and/or indirect influences on students'school experiences; their experiences, in turn, determine such outtomes of in-

.34 4 7

Envisioning Future Schools

terest for students as academic achievement, types of attitudes and behaviors,

such as vandalism, attendance, drop-out and the like.


Four studies' explicitly identified some aspect of the school's goals as an explana-

tion of differences in secondary-school effectiveness. Such goals included both

short and long-term outcomes considered important for students to achieve; they

also included the conditions in the school that would be necessary to accomplish

such outcomes. Exemplary secondary schools were characterized as having anunusually high degree of clarity about the purposes of schooling and the out-comes to be accomplished by students. These schools also emphasized cognitive

learning goals as especially important, within the larger purpose of providing abasic high-school education to all students. Effective schools make their goals

effective tools for decision-making: this was done by having written goal state-ments, using goals as the basis for communicating to others, insisting that

priorities fit goals, and using data to monitor progress toward goals and to refine

and redefine goals. Finally, such schools used their goals as a shared ideology to

provide a web of identification and affiliation which inspired loyalty to the school

among staff and students.As help.U1 as these results might be in helping shape a school-leader's vision

of the form and use of goals. they provide limited guidance in respect of the

substance of such goals. The broad social and professional trends discussed above

are helpful in this respect. In a nutshell, as Corbett, et al. (1990) suggest, schools

of the future will need to be both excellent and equitable. And whereas such

schools seem justified in continuing to emphasize cognitive learning goals, these

will not be narrowly defined by the conventional content of academic disciplines

or the basic intellectual skills measured by many current standardized tests. A

helpful illustration of the substance of goals for future schools is provided in a

selection of results from a series of working conferences we have held with cross

sections of the public and profession over the past two years (Leithwood andjantzi, 1989). Several broad goals, among many, emerging from these confer-

ences about which there is much consensus were, for example:

being able to reflect on one's actions and learn from such reflectionbeing curious and actively attempting to make personal sense out of as

much of one's world as possiblebeing able to manage information; to locate, sort, organize and assimilate

large bodies of information for the purpose of solving one's problemsassisting in the clarification of core family values and acting in a manner

consistent with such values (rcinforces moral and ethical standards)being able to cope with stress in a healthful manner

One of the most noteworthy features of these examples is the press they

create for future schools to consider how they might respond to the comprehen-

sive, developmental needs of the student viewed in the several contexts of

authentic life challenges; these include not only school and work but also family,


Developing Expert Leadership for Future Schools

community, and leisure ,ontexts. One might well argue that future schoolswhich were successful in addressing such a broad, complex array of goals wouldprovide much more meaningful experiences for a larger proportion of theirstudent-clients (perhaps especially the non-university bound student served par-ticularly poorly by today's secondary schools).


Nine studies` in our review of effective schools research identified qualities ofteachers found in exemplary secondary schools. These qualities addressed fouraspects of the teacher (several of which overlap with aspects addressed in asubsequent category labeled 'Programs and Instruction): relevant personal qual-ities, view of a teacher's role, disposition toward students and disposition towardcollaboration with other teachers.With respect to relevant personal qualities, teachers in exemplary schools

experienced high levels of job satisfaction and believed that they had employmentstatus. These teachers also demonstrated highly developed verbal skills and hadhigh attendance rates at school. Teachers in exemplary schools were dedicated tothe profession of teaching and expressed an ongoing interest in professionalself-improvement through taking courses, and the like. They accepted responsi-bility for a focus on student achievement, generally, fostering individual studcntgrowth, adapting school practices to meet individual student needs and forcreating additional learning opportunities for students. Such teachers were will-ing to meet with students informally outside school hours and took initiative indeveloping new teak tiing programs. They also modeled what Rutter et al. (1979)refer to as 'mental and moral probity' (for example, being prepared for class,being punctual, not wasting time in class and rarely ending classes early). Inexemplary secondary schools, teachers held high academic and behavioral ex-pectations for students which they transmitted to students in a positive way.They were sensitive to the age of the students, the peer pressures, and socialinterests associated with adolescence (this was especially important for teachers inintermediate and junior high schools).

Teachers in exemplary schools were disposed toward collaboration withtheir colleagues and others. This collaboration was evident in shared efforts todesign and prepare curricula, and participation in instructional improvementefforts, where they could learn from their colleagues. Collaborative attitudeswere also demonstrated by these teachers, through a willingness in their ownclassrooms to adhere to the school curriculum agreed upon by the school staff.Mutual observation and critique of instructional practices was standard practice inthese effective schools, as was participation with other teachers in setting andmonitoring well defined standards for teaching. Teachers in these schools estab-lished good relationships with parents and they actively encouraged interactionbetween parents and themselves.

The results of the research concerning teachers in exemplary secondaryschools is suniined up well in Lightfoot's comment (1983, p. 342) that 'Goodschools collect mostly good teachers and treat them like chosen people'. To theseresearch-derived qualities of teachers for inclusion in a school-leader's vision, the



Envisioning Future Schools

five social trends argue, as well for a larger repertoire of instructional skills, anissue explored in more depth in Chapter 7.

School Organizational Characteristics and Policies

Seven studies' identified school policies, affecting students and teachers as well asseveral other organizational features, as among the attributes of exemplary second-

ary schools.With respect to students, exemplary secondary schools emphasized academic

achievement through policies which reinforced high academic standards: theyalso enforced competency requirements. These schools minimized disruptivestudent behavior through the provision of simple, clear standards and rules forstudents, often established with student and parent participation, and by valuingand rewarding exemplary student behavior. Students' sense of affiliation with theschool was increased through policies which ensured students' access to counse-lors, provided students with a pleasant, comfortable environment, and ampleopportunities for students to take responsibility and participate in running theschool. This is related to the need, suggested by changes in families, for schoolsto take on a more nurturing role with students.

Policies of exemplary secondary schools focused on acquiring experiencedteachers (e.g., more than five years) and those with higher levels of attainment intheir discipline, ofien with advanced degrees. Such policies also include a systemfor monitoring teacher performance, usually through regular reviews of suchperformance by a school administrator, and discussions with the teacher concern-ing the results of such reviews. Exemplary schools, this research suggests, haveways of explicitly recognizing teacher accomplishments. Staff stability is encour-aged and resources are provided for ongoing staff development. Such activitytends to be more locally defined (as distinct from board or district defined) andmore group-based (as distinct from whole staff based). Teachers are assisted infinding the kind of staff development which they think they need. Given thepress toward greater instructional flexibility likely to be created by several of thesocial trends we have discussed (as well as one of the professional trends), staffdevelopment will take on ever increascd importance through the decade. Policies

in effective schools often provide teachers with access to discretionary funds tosupport their own improvement initiatives, a feature consistent with the initiative

to further empower teachers. These policies also involve teachers in school-levelpolicy decisions, often including, for example, the assignment of students toclasses. Duke and Gansneder (199(i) provide useful data on the complexities ofimplementing this policy in a way viewed as satisfactory by teachers.

Outside of policies focused specifically on students and teachers, exemplarysecondary schools tend to be relatively small in terms of student enrollment.They are organized to make the best use of available time, including maximizingthe use of hours in the school year, the design of a carefully structured schoolday, and flexibility for classroom-level decision-making within the day. Suchschools have considerable school-level discretion for determining the means to beused in addressing problems of increasing academic performance. This is consis-

tent with the professional trend toward decentralization. As McNeil's (l986)study points out, effective secondary schools are not primarily organized and



Developing Expert Leadership for Future Schools

managed to keep them running smoothly (although they do tend to run smooth-ly). Such schools are organized and managed to support the purposes of thecurriculum and the requirements for instruction implied by the school's philoso-phy of education. Effective secondary schools have the support of their schoolboards or districts for their improvement initiatives.

Programs and Instruction

Thirteen' of the twenty studies in our review identified some aspect of programsand instruction responsible for the exceptional impact of effective secondaryschools. For these purposes, the term 'program' included the curriculum contentpresented to students, the degree of choice among courses available to students,and the extent of 'articulation' among program components. As compared withtypical or ineffective schools, the programs of exceptionally effective schools hada strong academic curriculum, requiring more advanced coursework and moreeffort from students. A rigorous core curriculum, with limited elective choicesinvolving assigned homework, was also evident. These were core requirementswithin a 'rich' curriculum. The curriculum was carefully coordinated and organ-ized, the result of a preoccupation with coherence and integrity.

The term 'instruction' encompassed the teachers' general orientations toclassroom activity, the quality of specific teacher-student interaction, the use oftime for instruction, and the approach to collecting and providing informationabout student progress. Exceptionally effective secondary schoolswere associatedwith carefully planned instruction: this includcd a limited focus within a singlelesson, actively structured and directed classroom activities, establishment of aclear framework for pursuing sub-themes and small group or individual activ-ities; emphasizing opportunities for competence and achievement; and provisionof diverse experiences for students. Specific teacher-student interactions reflecteddefensible principles of learning. These included: clear communication of instruc-tional goals to students, promotion of extensive interaction in class, and use offactual questions to establish a foundation followed by higher order questionsrequiring more interpretation, for example. Efforts in exemplary schools studiedwere also made to keep classroom discussion vei y concrete and to demonstratethe relevance of curriculum material to the everyday world of students.

In these effective schools, efficient use was made of instructional time inmoving toward goals. This included: maintaining a fairly rapid pace oftenthrough teacher-directed forms of instruction, and minimizing time loss due toabsenteeism, lateness and inattention. A more rigorous, disciplinary climate wasestablished, as was frequent communication with students, either individually oron a whole-class basis. Systematic monitoring of student pcogress was commonin these schools. This included: assessments of performance keyed to instruction-al objectives, emphasizing students' being on task, keeping written records, pro-viding feedback to students, and rewarding students for significant achievement

These results reflect a set of goals much narrower than one is likely to aspireto for future schools. Less conventional forms of instruction need to b, incorpo-rated in a school-leader's vision, forms of instruction which rnake greater use ofother students and adults other than the teacher, as resources for learning.

385 1

Envisioning Future Schools

School Culture, Ethos or Cimate

Rossman, Corbett and Firestone (1985, P. 5) define culture as a 'unique set ofcore [norms], values and beliefs that are widely shared throughout the organiza-tion'. Rutter, et a/. (1979, pp. 55-6) use the term 'ethos' in reference to 'a climateof expectations or modes of behaving' suggesting that, in many cases, individualactions are less important in their own right than in the accumulated impact theyhave on what it feels like to be a member of the school organization. Although anabstract dimension of schools, effective schools' research has given culture, ethosor climate prominence as an explanation for differences among schools. Schoolculture is also an object for 'restructuring' in the profession on the grounds thatteacher empowerment and more decentralized decision-making is not likely inthe traditionally isolated culture of most schools (Feiman-Nemser and Flodden,

1986).Results of our review of eleven studies' suggest that exemplary secondary

schools are safe and orderly. Neither staff nor students are concerned aboutphysical harm; the school is free of discipline and vandalism problems; andteachers understand and are not threatened by adolescents. The school refuses tocondone or conceal indiscipline, and staff have what Lightfoot (1983, p. 342)refers to as a 'fearless, empathetic regard for students'. Cultures in exemplaryschools are positive: there is less emphasis on punishment and critical control,and more attention to praise and student reward, especially with regards to theirindustry and exemplary behavior. Student self-control is encouraged, and bothstudents and teachers take school seriously.

Teachers engage in precise, concrete talk about teaching practice and itsiniprovement in exemplary schools, and specific support is provided for discus-sion of classroom practice. As already suggested, the shared technical cultures ofexemplary schools are built on norms of collegiality, collaborative planning, andcontinuous improvement. The staff are cohesive and have a strong sense ofcommunity. Norms are established which contribute to the integration andcohesiveness of the student body; they may be done through collective events ofvarious types that arc directed so as to reinforce the purposes of the school (e.g.,scholastic competition). There is reciprocity between and among staff and stu-dents. These cultures are also student centered. This is evident in much of whathas been said about the school culture already. Schc sl-wide recognition ofacademic success is promoted.

School-Community Relationships

School-community relationships were identified, in three studies.' as an impor-tant discriminator among schools which varied in effectiveness. Unlike the case

in elementary schools, such relationships were not with parents directly but wereinstead with non-parents who had a direct contribution to make to the school,and with the community at large. Effective secondary schools, it was reported,

made effective use of such con,munity resources as volunteers and student tutors.Solid working relationships were developed with local business and industry for

career training, for example. Similar types of relationships were also evident with

colleges and universities for assistance to academically talented students. These


Developing Expert Leadership for Future Schools

schools were responsive te their particular social and political miheux, andgenerated high levels of community support. These characteristics reflect quiteliterally the professional trends, discussed earlier, regarding school-communityrelations: they also reflect a concern for public accountability. We might add towhat these schools arc doing, the establishment of closer links with social serviceagencies in order better to provide the social capital that some changing familystructures are not providing.

ElementarySeondary School Difjerences

It has been suggested that effective secondary schools are likely to differ fromeffective elementary schools for a variety of reasons (Farrar, Neufeld and Miles,1984, Firestone and Herriott, 1982). As compared with elementary schools,secondary schools are usually larger, have greater role differentiation, and pursuemore diverse outcomes. This makes communication more difficult, complicatesthe process of arriving at a consensus about instructional goals, and reduces thepossibility of-principals exercising direct instructional leadership. Murphy andHallinger (1985), however, argue that these differences are more apparent thanreal in terms of their consequences for effective schooling. As a way of inquiringabout elementary-secondary school differences, the twenty secondary schoolstudies were compared with five reviews of research on effective elementaryschools (Cohen, 1982, Duckett, 1980), Edmonds, 1979, Mackenzie, 1983, Weil,et al., 1984). This was a 'convenience' sample of reviews, for which no specialjustification is ofkred. Our analysis suggests that, as compared with effectiveelementary schools, effective sccondary schools may:

pursue a broader range of goals (providing a basic high-school educationvs. basic skills, for example)be more concerned about developing a sense of community and affilia-tion within the school, and use goals, in part for that purposeattribute more importance to the job satisfaction, employment status,verbal skills, and attendance rates of teachersattribute more importance to such basic beliefs of school-leaders asviewing teachers as professionalsrequire school-leaders to consider a broader array of factors in the schoolin order to exercise influencehave to address problems related to size of staff and student body moreexplicitlyrequire more school-level decision-making discretionexpend more effort on the design of a program which is useful for allstudents and provides enough variety to address a more diverse set ofneedshave to promote and support more precise, concrete talk amongteachers, concerning their classroom practiceshave less need for close parent involvement

In spite of such real or apparent difkrences, the comparative analysis lendssupport to at least some aspects of twenty-three of the thirty-four characteristics


5 3

Envis'oniq Future Schools

of effective secondary schools identified in the review of the twenty original

empirical studies.


This review of research concerning six dimensions of effective schools illustrates

how such evidence might be used to envision a future school. In the case of most

dimensions, the research alone had to be supplemented with inferences drawn

from one or more of our broad social trends. What is remarkable, however, is

the extent to which the features of today's exemplary schools, reported in the

research, incorporated features which anticipated many of the consequences for

schools of those broad social trends which were considered. They also appear to

have implemented a number of thc professional trends being advocated for

schools just now. These results, then, support our initial contention that research

about exemplary practice in the present is a useful resource for envisioning future

schoois, assuming a not too distant horizon.In sum, the chapter was designed to assist school-leaden in developing a

defensible vision of future schools. At first blush, the attempt to offer this

assistance (especially in such limited space) seems naively ambitious. There is,

after all, a wealth of information, relevant to such a vision, to which wc did not

refer. What we hoped to accomplish was not a definitive vision for school-

leaders. That is neither possible nor desirable. Rather, we wanted to demonstrate

that there are specific sources of information to be used in developing a defensible

vision. We also wanted to demonstrate that developing a defensible vision entails

careful, systematic effort; some aspects of what makes a good school are suf-

ficiently clear (from research evidence, for example) that they ought to be

carefully considered in any vision; and in order to provide real guidance in future

problem-solving, a vision must be quite detailed. A blurred vision offers little

guidance for school-leaders and their colleagues.


1 Included were Huddle (1986). Lightfoot (1983). Ford Foundation (1984) and Lipsaz

(1984).2 See Huddle (1986), Murphy and Hollinger (1985). Roucche and Baker (1986). Lipsitt

(1984), Rutter, et al. (1979), Ford Foundation (1984), Goodlad (1984), Madaus, Kel-

laghan and Rakow (1976), Lightfoot (1983).

3 Studies providing data concerning these issues included Harmsch (1987), Huddle

(1986). Ford Foundation (1984), Lipsitz (198), McNeil (1986), Goodlad (1984) and

Rutter et al. (1979).4 These studies included Lipsitz (1984), Roueche and Baker (1986). Huddle (1986),

Lightfoot (1983), Coleman and Hoffer (1987), Keith and Page (1985). Arehart (1979).

Goodlad (1984), Walberg and Shanahan (1983), Harmsch (1987), Murphy and Hollinger

(1985), Morgan (1979). Frederick, et al. (1979).

3 See Lipsitz (1984), Roueche and Baker (1986), Huddle (1986), Murphy and Hollinger

(1985), Goodlad (1984), Rossman, Corbett and Firestone (1985). Rutter, et al (1979).

Coleman and Hoffer 0987). Lightfoot (1983). Ford Foundation (1984), Gunn and

foldaway (1986).6 See Huddle (1986). Ford Foundation (l'IS4). I Tsar (1984).



Chap ter 4

Expert School Leadership on theHigh Ground and in the Swamp

In the varied topography of professional practice, there is a high hardground overlooking a swamp. On the high ground, manageable prob-lems lend themselves to solution through the application of research-based theory and technique. In the swampy lowland, messy confusingproblems defy technical solution. (Schön. 1987, p. 3)

In Chapter 1, we made a case for thinking about the leadership process asproblem-solving. A problem, as Baird (1983) and others define it, has threeIngredients: givens the current fact or situation (e.g., this parent standing in myoffice complaining aboot his son's teacher); goals - different, more valued fact orsituation (e. g. , this same parent satisfied with his son's teacher and not standingin my office); and obstacles or constraints that must be overcome before thegiven state can change (e.g., the parent's perception of the unfairness of theteacher's grading practices). Research on problem-solving (e.g.. Fredericksen,1984, Shulman and Carey, 1984) has identified two sources of variation in thedifficulty experienced by people in solving problems. One of these sources ofvariation is the clarity or extent of knowledge a person has about those threeingredients of a problem; the other is the specificity of the steps the person is ableto identify in order to successfully overcome obstacles or constraints, and therebytransform the complaining parent into the satisfied parent.

When schoul-leadcrs are well informed about the givens in a problematicsituation, clear about what goals, if accomplished, would solve the problem, and(perhaps through previous experiences with similar problems) have reliable pro-cedures for overcoming obstacles, they are on what Schön (1983) refers to as the'hard high ground'. But the ground on which school-leaders base their practicesbecomes increasingly swampy, as fewer and fewer of these conditions are met:the swamp is especially deep when one only vaguely understands the presentsituation, has no clear way of knowing what would be better, and lacks proce-dures for addressing the obstacles or constraints in the situation. The mostpredictable aspect of the future tor school-leaders will be the continuing need tosolve swampy problems.

This is our understanding of at least part of what Vail (1989) means in hisreference to leaders of future organizations being 'in permanent white water' (aswamp with whitecaps!!). No matter how complex a problem might be, in anobjective sense (e.g., technically difficult or fraught with interpersonal conflict).


S:hool LeaderAp ri the High Ground and in ihe Swamp

its `swampiness' is a purely suijective matter. So, the more a person knowsabout how to solve any problem. the less swampy it becomes for that person. Atfirst blush, then, it seems that given enough time and effort, we ought to be ableto dramatically reduce the number of problems. faced by school-leaders, that arctruly swampy. More research, better education of those entering the role, grow-ing experience in the role, and ongoing professional development seem likeobvious strategies to such an end. But this picture of a gradually draining swampassumes both a static context, within which future school-leadership will beexercised, and an unchanging set of expectations for those exercising such leader-ship. No one in their right mind would bet on that assumption. Indeed, thenature and rate of change that can be anticipated in the future means that some ofthe problems, now on the high ground, will become largely irrelevant fortomorrow's school-leaders. This will be the case, for example, because of g..ow-mg availability of 'user friendly' technical solutions (like, for example, construct-ing school timetables), because altered professional norms will support practiceswhich, at present. remain tough problems to deal with in many schools (e.g.,helping some teachers appreciate the value of genuine peer collaboration), as wellas for many other reasons. Indeed, if predictions about exponential rates ofchange in the future prove to be true, then even a substantial increase in efforts to'dehydrate' future school-leaders' problems may be accompanied by higher waterlevels in the swamp.

All of this is to say that expert school leadership will depend in the future, asit does now, on dealing effectively with both high ground and swampy prob-lems. Chapters 5 and 6 describe the nature of such expertise. The next twosections of this chapter address three sets of questions, each with a concern aboutthe implications for future school leadership:

What basic characteristics of problems cause them to be considercd,by current school-leaders, as high ground or swampy?What specific types of problems do school-leaders presently classifyas high ground and swampy?How do school-leaders decide on the importance of a problem?

The primary data used to address the three sets of questions in the remainderof tnis chapter are drawn from five studies. Three of the studies were about howgroups of expert Ind non-expert elementary principals (Study One: Leithwoodand Stager, 1986), secondary principals (Study Two: Leithwood and Stcinbach,1990) and, for comparison purposes, superintendents/chief education officers(Study Three: Leithwood, 1988) classified and managed their problems. Twostudies inquired about the nature of both elementary and secondary principals'problems, and the ratio of high ground to swampy problems (Study Four:Leithwood, Cousins and Smith, 1989;90, Study Five: Leithwood, 1990); thesestudies also exemined the frequency of occurrence of problems over a schoolyear.

What Makes a Problem Swampy for School-Leaders?

To this point, a swampy problem has been defined as one about which the solverpossesses very little knowledge, regarding how to accomplish sonic valued goal.



Developing Expert Leadership for Future Schools

We have also said that this is a highly subjective matter. One school-leader'squagmire may be another's dance floor. It depends on their existing knowledge,skills and dispositions presumably a function of such things as previous experi-ence with similar problems, relevant education and support. This is limitedclarification, at best, of the meaning of a swampy problem, however. Thequestion to be addressed in this section is: Givt.n the unique demands andexpectations of school-leaders, what causes them to classify a problem as difficultor swampy?

In Study Two (Leithwood and Steinbach, 1990), secondary principals, as agroup, were readily able to indicate their criteria for classifying problems, butthere was not much consensus among them. Impact on staff (the extent to whicha problem affects individual staff) was the criteria identified by the largestnumber of principals. At least two principals, out of the eleven in this study,identified the remaining as: availability of clear procedures; number of peopleinvolved in a solution; the possibility of value conflicts arising; whether thesolution would bc acceptable to all (a 'win-win' solution); and whether thoseaffected by the solution would accept it (be 'reasonable', even if the solution didnot favor them). The extent to which the problem could be solved within theschool, and therefore, be 'controlled' was also a basis for estimating problemdifficulty. As one principal said:

[easy problems] are in-house we have control of the resources and thepeople being involved. It becomes more difficult as you move out [ofthe school].

Data from elementary principals in Study One was similar, in respect tothree of the seven criteria: availability of clear procedures; impact on staff (essen-tially, problems affecting people directly and likely to have significant emotionalcontent); and number of people required to solve the problem. On the otherhand, CEOs in Study Three used all of the same criteria as secondary principals,in their estimates of problem difficulty. The extent of a problem's impact acrossthe school or school system was a criterion assoctated with problem difficulty, byelementary principals and CEOs, but not secondary principals.

What are we to make of these data? Problems tend to be considered swampyby expert school-leaders, the more that they are primarily 'people problems', aresult also evident in Goldman et ai.'s (1990) study. Problems have this character,not only when they are directly about a person, as in the case, for example, of theperformance of a teacher. They also have this character, as more people arerequired for a solution, as the particular people involved are viewed as inflexible

perhaps a function of significant value differences relevant to the problem, andas the problem or its solution impacts on increasing numbers of people. Incontrast, problems are not seen to be swampy when there is a clear procedureavailable for solving it. Most of the time, such problems are relatively technicalm nature and have been encountered routinely in the past. For example:


leasy problems are] things like budget allocation in as much as I havedone it for a long time. I know how to negotiate with the people whoare responsible thr the various accounts.

5 7

Expert School Leaders lup on the High Ground and in the Swamp

Such problems can also be solved by those within the organization, often thoseover whom the principal has some vested authority.

While the designation of a problem as swampy is, theoretically, a purelysubjective matter, the common features of school-leaders' work (e.g., state,district and school contexts; see Goldman et al., 1990) mean that many identifythe same problems as swampy or not swampy. In our research, for example, aproblem almost always designated as swampy by school-leaders, from among anarray given to them, was the Principal-Entry problem:

You have been assigned to a new position. The present principal, who isvery highly regarded by staff, community, and students is being movedto a larger school after only two years in this present assignment. Theschool community (staff and parents) arc very displeased that a newprincipal has been assigned. They feel that the board has not consideredtheir wishes. How would you enter this situation?

But our setting-school-objectives problem was almost always considered to he ahigh ground problem:

Your new school is one in which staff have never been involved in thesetting of school objectives and are not apparently very interested indoing so. You have conie to believe that it is a very important thing forstaff to set school objectives and ro evaluate them at the end of the year.

A comparison of these two problems illustrates, a bit further, the meaning ofthe results described to this point. The Principal-Entry Problem involves manypeople; there is the potential for value conflicts; emotions are intense; and theentering principal's reputation is on the line. Furthermore, one of the mostprominent groups of people in the problem, parents, owe no bureaucratic alle-giance to the school organization. Setting school objectives also involves theparticipation of many people, but it is the school staff who are involved. Mostexpert principals are also able to identify an array of well-developed proceduresfor solving this problem and have had experience in using one or more of theseprocedures at least several times.

In sum, then, the more that a problem involves people, and requires thecooperation and participation of increasing numbers of diverse people, the swam-pier it is likely to be for school-leaders. This, of course, places a high premiumon the quality of one's interpersonal skills as a prerequisite for solving swampyproblems in schools (the 'considerative' aspect of future school-leaders' practicesdiscussed m Chapters I and 2).

Wilat Specification Kinds of Problems Do School-Leaders FindSwampy and How Prevalent are They?

Even though swampy problems arc, by definition, challenging to solve, much oftheir significance in the work of school-leaders depends on how often they areencountered. If swampy problems are infrequently experienced, one might argue

Developing Expert Leadership for Future Schools

that it is not especially crucial to devote much time or effort to further develop-ing school-leaders' capacities for solving them as effectively as possible. Mud-dling through will do. But swampy problems are expensive to deal with: theyrequire a lot of thought. And school-leaders do not have time to give a lot ofineffective and inefficient thought to many problems. So, what is the case? Areswampy problems a common occurrence or not, and, specifically, what are they?

Studies Four and Five (Leithwood, Cousins and Smith, 1989/90; Leithwood,1990), including a total of twenty-six elementary and twenty-six secondaryschool-leaders, provide the only evidence that we are aware of concerning thisquestion. So the answer must be considered tentative. Relevant parts of thcsestudies documented the full array of problems reported by principals over aneight-month period, including the principals' designation of each problem, aseither relatively routine or non-routine (swampy). These problems were organ-ized into sixteen categories, frequencies of occurrence calculated and com-parisons made between elementary and secondary schools, and routine andnon-routine problems (Study Five r,!sults replicated those of Study Four almostexactly).

Four problem categories were reported much more frequently than the rest(these frequencies arc drawn from Lcithwood. Cousins and Smith (1989/9)))


I. Teachers (247 Problems)

I. Assignment of Teaching Duties2. Conflicts Among Teachers3. Conflicts Between Teachers/Students/Administration4. Curriculum Review, Development, Implementation5. Dereliction of Duty (reporting, deadlines, supervision)6. Dress Code7. Extra-Curricular8. Judgment of Teacher-Proposed Ideas9. Level of Competency

1(1. New Teachers11. Personal Problems12. Professional Development13. Staff/Department Meetings14. Teacher Coverage15. Teacher Evaluation16. Teacher Exchange

2. School Routines (1 38 Problem))

1. Assemblies2. Attendance3. Budget4. Commencement PlanningS. Dances6. Drills and Routine For Students7. Feeder-School Visit8. Field Trips

Fire Drills

5 9

Expert School Leadership on the HO Ground and in the Swamp

W. Fund Raisers11. Graduation Awards12. Home Room Visits13 IPRC (Individualized Program and Review Committee) and Special

Education Meetings14. Public Address Announcements Meeting15. Re-timetabling of Classes16. Registering Students17. Report Cards18. September Report19. Student Council Meetings/O. Teacher Routines/Plans21. Timetabling22. University Night

3 Students (113 Problem.,

I. Abuse2. Adult Students

Attendance4. CafeteriaF. Commendation6. Complaints7 Discipline8. Evaluation9. Injuries

In Placement11. Special Requests12. Student Council13. Student Problems14. Vandalism


4. Parents (105 Problems)

1. Communication2. Complaints3. l'arent Councils/Groups4. Parents Night5. Parental Involvement in the School or lack thereof ...

As the specific problem labels within each of the fOur categories indicate fully,two-thirds of principals' problems (402) revolve around the internal workings ofthe school, its staff and clients. These are problems over which the principal has afairly high degree of control. The remaining problems arise from aspects of thcinternal workings of the school which appear to require very infrequent attentionby the principal (e.g., non-teaching staff plant. special events). Requiring re-latively infrequent attention by principals are problems arising from sources alsoexternal to the school. Consistent with evidence about external influences onprincipals, reported in ChaNer 2, senior administrators (72) are thc most fre-quently & ited of thcse sourL cs. Lhey place accountability dvmands on principals.



Developing Expert Leadership for Future &hoots

visit their schools, provid: approval or non-approval of principals' initiatives,request attendance by principals at board meetings for a variety of purposes, andinsist on adherence to system procedures. Trustees, the Ministry of Education,and outside agencies of several types (e.g., social-service groups, community-health groups) appear to impinge very little on principals' problem space.

Boyd and Crowson's (1981) review of research concluded that principalstypically have an 'insider' focus, and spend the bulk of their time on organiza-tional maintenance tasks and pupil-control tasks. Our evidence concerning thetypes of problems encountered by principals is consistent with only a part of thisconclusion. Three of the four most frequently encountered categories of prob-lems were found inside the school: teachers, school routines, and students. Thefourth, parents. could also be viewed as an 'inside' problem category. If parentsare so considered, 'outside' problems encountered by principals amounted toonly about 19 percent of the total.

It is not difficult to understand the need for school principals (indeed middlemanagers in other organizations, as well) to have an 'insider focus'. This whythey were hired, one may argue. But is the focus of their work actually justmaintenance and control? This question is best answered by examining specificsubcategories of problems. Such problems do not suggest a necessary preoccupa-tion with maintenance and pupil control. For example. the subcategory of'teacher' problems entitled 'assignment of teaching duties' contained fifteen citedproblems in Study Four (not shown in text). Several of these problems clearlywere of a maintenance nature (e.g., plan for lunchroom supervision, findingsupply teachers). But most could bc plausibly linked to the school's instructionalprogram (e.g., setting goals with teachers, reorganizing thc grade 7 class, arrang-ing more planning time for teachers). Over the total of 247 specific references inStudy Four t., 'teacher' problems. at least a majority had the potential for a directimpact on instruction. Further, 'student' problems were by no means limited tothe control of students. The majority had some direct link to the likelihood ofstudent growth: for example, incidences of child abuse, counseling adult studentson diploma requirements, 'behavioral' student running away from school Alarge minority of this category of problems did involve control: discipline,attendance, and the maintenance of order were examples. Our data suggestingthat an Insider thcus by school-leaders is likely. but that such a focus may wellconcern school improvement, not simply maintenance matters, is consistent withevidence from the work of effective principals, reported by Smith and Andrews(1989) and by Pavan and Reid (1990).

In terms of categories of problems experienced, there appeared to be fewdifferences between elementary and secondary principals, with respect to non-routine problems. 'Student' problem% were .he sole exception to this resultelementary principals encountered over 50 percent more of such problemsSubstantial differences were evident with respect to four categories of routineproblems and in each case. elementary principals facing a greater numberapproximately three times as many 'parent' problems, eight times more 'plant'problems. fourteen times more problems associated with other principals, andtwice as many school-routme problems.

In hue with earlier discussions of the criteria school-leaders use in determm-nig problem difficulty. the incidence of non-routine problems is much higher inthree of the four categories in which principals encounter most problems overall


6 1

Expert School Leadership on the High Ground and in the Swamp

students (43 percent), teachers (1/ percent), and parents (19 percent) - all 'people'problems. While problems related to school routines are frequently encounteredby principals, their responses appear to bc well rehearsed, requiring little con-scious attention. Only three non-routine problems were reported in this category(these results conflict, to some extent, with data reported in Chapter 10, indicat-ing that school-leaders do not feel well prepared to solve such problems).

Studies Four and Five (Leithwood, Cousins and Smith, 1989/90; Leithwood,1990) also reported ratios of routine to non-routine problems in each of sixproblem categories. These ratios are a much better estimate of just how'swampy' the problems, are in each category, for principals. At least this is thecase where the data provided a sufficient sample of problems, for such estimates tohe meaningful. 'Student' problems are clearly the most non-routine, in the mindsof principals. Problems related to the school plant are the next most non-routine,but very few such problems were reported. Categories of problems including thecommunity at large, the Ministry of Education, parents, teachers and vice-principals/department heads were similar in their ratio of routine to non-routineproblems. There was about a five to one ratio of routine to non-routine problemsencountered in these five categories, as well as in all sixteen problem categoriescombined

Descriptions of the day-to-day activities of principals have pictured theirjobs as hectic, fast-paced. characterized by brief encounters, and spontaneousface-to-face interactions (Widower and K metz, 1982. Martin and Willower, 1981,Wolcott. 1978, Morris. et al.. 1)84). Such characterizations of what principals do,often leave thc impression that their problems, while numerous, are largelyroutine. In contrast, our data suggest that experienced principals perceive a muchhigher proportion of their problems to have, at least, significant componentswhich are non-routine (about one in five and considerably higher in the case of'student' problems). Principals do not see themselves simply applying a well-rehearsed repertoire of solutions, over and over again, to the same problemstechnical view of their role. Rather, adaptation of old solutions to new contextsand circumstances. as well as fresh thinking about largely novel problems seemto describe better a significant proportion of the demands faced by principals.

This haracterization is consistent with Schiin's (1083) depiction of expertpractice in other prokssional fields. It is also consistent with what we knowabout the number of reforms. new expectations, and shifting environmentsschools now face. Furth -!more, the willingness and ability of principals to seenovelty in problems which have a familiar cast appear to he important featuresof expert school-leader problem-solving. For example, in comparing a sample ofnon-expert and expert principals, Leithwood .oid Stager (1089) found that non-expert principals were much more likely to become hostages to their existingknowledge and experieme; they were unable, as a result, to recognize newfeatures of a problem that required special attention if the problem was to beadequatel y resol ved Expert principals were quick to sec and to act on suchfeatures.

Studies reviewed in Chapter 2 suggested that principals perceive senioradministrators not only as A sigmticant source of problems, but as an impedimentto, rather than a resource for, their school-improvement efforts. For example.( rowson and Morris (1085) found that from a third to a half of Chicagoprincipals' school-site activity was governed Fy what they called 'hierarchical


Developtng Expert Leadership for Future Sdwols

controls'. Principals, in Leithwood and Montgomery's (1984) study, reportedthat the major hurdles they faced in improving their schools were: hierarchicalstructures which made change difficult, excessively rigid and time-consumingdistrict policies and procedures, provision of inadequate resources, and the con-servative stance of central administrators toward school-initiated change. Similar-ly, Duke's (1988) inquiry about why principals quit their jobs suggested fivesources of job dissatisfaction, three of which were associated with the work ofsenior administrators.

Available evidence, although quite limited, appears to reinforce thc claimthat school-leaders often view the work of senior administrators as less thanhelpful. Senior administrators were viewed by school-leaders in Studies Four andFive as the greatest source of problems outside the school (about 13 percent of thetotal). Yet, less than 1(1 percent of the problems identified were considered to benon-routine by school-leaders, and even these included a number with a clearmaintenance focus. Problems one might have expected school-leaders to view asnon-routine, such as development of a professional growth plan, were not. Thissuggests that at least some procedures characteristically relied on by senioradministrators as change strategies may turn out to be relatively benign as theya re implemented.

In brief, the limited evidence which is available suggests that a surprisinglyhigh proportion of school-leaders' problems are non-routine a higher ratio forsecondary than elementary leaders. Most of these are 'people' problems, morespecifically people within the school. The most noteable exception to the 'insider'focus concerns problems related to district administrators.

How Do School-Leaders Decide on the Importance of a Problem?

Time, according to most school staffs, is the scarcest of all the resources whichthey need. And since (like land) no one is making any more of it these days,decisions school-leaders make about how to allocate their time centrally, affectthe contribution they are able to make to their schools. No school-leader hasenough time to address carefully all the problems encountered in the course of aday. Variations, among school-leaders, in their ability to pick the 'right' prob-lems to address seriously, the 'right' problems to ignore, and the 'right' problemsto pass over lightly, account for much of the variation in their effectiveness.Evidence regarding this aspect of school:leaders' problem-solving was providedby Studies One and Two (Lenhwood and Stager, 1986; Leithwood and Stein-bach, 199(J).

Results of these studies suggest, first of all, that the most direct response totins question of prioritizing for non-expert school-leaders was that thcy don't.That is, moderately effective school-leaders appear not to have an explicit processfor deciding which to address (except the order in which problems are encoun-tered). As one such person said, when asked about his way of prioritizingproblems.

To be honest with you. I really don't. I don't think of things and try topm them in slots. I try to deal with a problem when it conies and I don't


Expert &Iwo! Leadership on the High Ground and in the Swamp

believe in delaying. If somebody is really upset about something. I wantto know right then and there. I don't want to leave it for half an hour, orgo home and think about it.

In contrast, expert school-leaders quite consciously use an explicit strategy (otherthan order of encounter) for sorting, in their daily problem-solving. For example:

Yes. I do. I'm fairly conscious of that ... I have a VP in thc school andshe knows where I'm connng from, and I ask her to help keep me ontask. Last night at the divisional meeting we were reviewing our reportcards. I outlined some of the modifications wc have made ... and thatwas strictly a 'review-tell' decision. Now I then moved into an area ofevaluation ... [cites an example of problem and process used] ... Thatwas getting into a somewhat of a 'sharing' but it was kind of mixedbecause I was laying on what I think is somewhat of a base from whichto work.

Beyond the presence or absence of an explicit problem-sorting strategy,expert school-leaders consider a number of other criteria when assessing thepriority a problem ought to get for thc time available. Some of those who sortproblems according to who will be involved in them do such sorting accordingto organizational groups (e.g., 'tile it for the lead-teacher meeting'. 'file it for thestaff meeting'); others sorted in tcrms of type of decision (e.g., 'tell' vs. 'sell', vs.'share), role responsibilities (e.g., a 'teacher-owned' problem). or numbers in-volved (e.g.. joint problem vs. individual problem). The main point seems to bethat these school-leaders have some way of consciously discriminating amongproblems and that they consistently apply this to assist in their problem manage-ment. Their moderately effective counterparts do not.

Most school-leaders are concerned about student and staff needs, and feeloverburdened with paperwork. identifying this as something prev,:nting thcmfrom spending as much time as they would like on other priorities. Expertschool-leaders tend to award priorities to problems impacting on the schoolprogram and on overall school directions (without ignoring the problems ofindividuals). When asked about problems which he prioritized, a participant inStudy Two said:

ones that are going to have great impact on people, in terms of coopera-tion. A problem that impacts on one person versus one that impacts onthe whole school, I see quite a difkrence in approach to that, in terms oftime and the amount of communication in,'olved because you're dealingwith so many people.

Non-experts, on the relatively infrequent occasions when they establish prioritiesamong problems, tended to do so on the basis of which group of people isinvolved. As an example of this, one stated:

any problem that is going to affect the kids get priority. Then, after that,tf it affects staff ... and then head office gets to the bottom, unless headoffice is saying 'Get the report in by three days tune!'

Developing Expert Leadership for Future Schools

Some non-experts also set priorities in terms of deadlines, order of arrival, oramount of time required for solution. One moderately effective school-leader inStudy Two said: 'No, I can't say that I have any way of deciding what gets donetirst.

As compared with non-experts, experts have more deliberate strategies formanaging their time and for ensuring that time is available for high-priorityproblems. For example:

1 have to come in about 7:15 to deal with my items and there is a chunkof the day that I can also work on my items once the programs arestarted.

However, these strategies of experts extend beyond daily time-managementroutines. One expert elementary school-leader described the importance of pre-dicting what the problems will be and thus preventing their occurrence. He notedthat many 'crises' are the result of not being prepared:

when that happens. you can find that you are spending most of yourtime I during a term I dealing with it because usually if one thing hasgotten out of whack, so has another ... Usually during that time, whileyou're kicking yourself, you're saying 'That's not going to happen againnext term!' You are preparing yourself and you are predicting.

This approach to time management may also he viewed as a problem-preventionstrategy. Differences between non-experts and experts in problem-preventionstrategies such as these are pronounced. Much more than the non-experts,experts are explicitly concerned with preventing unwanted problems from arisingin the first place, and have more strategies to do so. For example, they are out oftheir offices, around the school and in classrooms regularly and frequently; thisallows them to detect the early signs of discipline problems, instructional prob-lems and the like. They attempt to focus staff on program initiatives that areexciting and absorbing, thereby reducing the likelihood of 'staff morale' prob-lems. Most of these experts, furthermore, arc quite aware that developing excel-lent programs and instruction not only contributes to student growth. but alsoprevents many parental complaints from arising.

Summary and Implications


We view the generic process of school leadership as problem-solving where thenotion of a 'problem' is frcc of the negative connotations with which it is oftenassociated in everyday talk. For our purposes, a problem is any challenge or taskor job to be done which requires thought (conscious or unconscious) and sometype of action. This chapter has distinguished between two types of basic prob-lems facing school-leaders labeled, metaphorically, 'high ground' problems and'swampy' problems. High ground problems are those about which a school-leader possesses a considerable amount of relevant knowledge, and probably has



Expert School Leadership on the High Ground and in the Swamp

had considerable previous experience. Because of this background, the school-leader's response may be reliably effective, relatively smooth, largely automatic,and consume very little conscious, thoughtful effort. In contrast, problems areconsidered to be swampy when one possesses little relevant knowledge andprobably has little previous experience in resolving at least some non-trivialaspect of the problem. Problems of this sort require the invention of responseswhich may need to be fine tuned, in process. Above all, such problems demandconsiderable thought.

The distinction between high ground and swampy problems is one of degreenot of kind: more like taking a humidity reading than sorting into two piles.Nevertheless, the distinction proves to be quite useful in better understandingexpert school leadership. Furthermore, the distinction is entirely a subjectivematter. The existing knowledge, skills, and dispositions of the individual school-leader will determine the degree of swampiness of a given problem. And it is

because increases in the individual school-leader's repertoire change the humidityreading of a problem, that swampy problems are eventually allowed to dry out.

Notwithstanding the subjective nature of these problem designations.school-leaders' experiences and resulting repertoires share common features. Thisexplains why the research reported in this chapter found similarities in thc typesof problems considered by school-leaders to be high ground and swampy.Problems were considered swampy when they involved many people. especiallyincluding people outside ;he school whose responses could not be controlled inany systematic w ay. Problem% were thought of as high ground when they wereof a more technical nature, where a well rehearsed procedure for solving wasaVadable.

The chapter also examined evidence which suggests that swampy problemsare a surprisingly high proportion of the problems encountered by school-leaders; the proportion may be as high as onc in live for secondary school-leadersand only slightly lower for those iii elementary schools


Evidence reviewed m this chapter has at least six implications for developingexpert leadership thr future schools. First, we interpret the evidence about thenature of school-leaders' problems to suggest that the 'naturally' occurring prob-lems encountered by school-leaders provide them with ample Opportunities threxercising leadership. But this leadership is not often the type that comes to mindwhen the term 'instructional leadership' is used; the term implies direct andpervasive interactions with teachers bout their teaching strategies Our datasuggest that. normally, expert school-leaders influence instruction by establishingthe conditions within whit Ii stall instrut lion occurs. In our view, this is especial-Is mniportaim v ork thr thture school-leaders if the teaching profession continuesits current trend toward accepting more responsibility for instructional improve-tnt nt Creating conditions to enhant e the instructional improvement elthrts offiat hers is something school-leaders in positions of formal authority are wellsituated for. (Chapter 9 elaborates on the importance of school culture, aS'tondition which critically frames the instructional practices of teachers and

h can be sionficantly shaped by school-leaders

5 i


Developing Expert Leadership for Future Sthools

Under circumstances of increased teacher responsibility for instructionalimprovement, there will be less and less need to assume that unless school-leadersare constantly in classrooms observing instruction, they have little effect on thequality of education in their schools. School-leaders, in the future, will beconfronted, as they are now, with literally hundreds of small, spontaneouslyoccurring leadership opportunities (problems to be solved). The key to expertleadership will be to have these opportunities accumulate in a consistent anddesired direction. Many current school-leaders are unable to develop such con-sistency, most likely, we believe, because they do not have a set of goals, values,and vision for their staffs and schools, clearly formulated, at least in their ownminds. This further reinforces the importance for future school-leaders of de-veloping a widely shared, defensible vision, as we said in Chapter 3.

A second implication for developing future school-leaders, emerges fromdata concerning the proportion of non-routine to routine problems experiencedby school-leaders. There is a large enough proportion of school-leaders' non-routine problems to suggest that experiences which are effective in improving thecapacity to solve such problems would be very useful. There is, as yet, littleattempt to focus school-leader development explicitly on such capacity, and littleguidance about the nature of experiences that would be suitable. This is animportant area taken up much more carefully in Chapter 12.

Third, our results also argue strongly for the development of a betterunderstanding of how the work of district leaders can foster development ofexpert school leadership. The recent focus on effective schools has often createdthe impression that the larger organizational context in which schools function isirrelevant to their decisions. But evience reviewed in this chapter supportsreviews of recent multi-level research (Leithwood and Jantzi, 1989) suggesting anurgent need to better understand how school districts and senior administratorscan best facilitate the work of school... their lea lers. This matter is addresseddirectly in Chaptet 15.

Fourth, tlic secondary as compared with elementary-school administrativecontext that emerges from this study is one which is marginally less fast-paced,with increased opportunities, and a greater need to focus one's energies on alarger proportion of non-routine problems. At the same time, the role seemslikely to demand much more sys.ernatic and sustained attention to the develop-ment of staff, such as departmeat heads, who mediate principals' relations withteachers and studznits. Without such ,:stention, it is easy to imagine the secondaryprincipals influence on studei.ts to be undetectable. Our own informal butsubstantial experience suggests, however, that a great mar:y secondary principalsfeel quite uncertain about how to proceed widi such staff development. Thedevelopment of specific. credible advice on this matter would be very helpful.But such advice requires, as a prerequisite, a much better understanding of thesecondary department-head's role than is presently available.

Two final implications for developing future school-leaders are evident inthe research reviewed in this chapter. It will be important to help such leadersacquire defensible procedures for setting priorities among their problems, and towork especially well with niany different types and groups of people.


6 7

Chapter 5

Expert School Leadership on theHigh Ground

Expert school leadership on the high ground depends upon the application of

significant amounts of knowledge relevant to the problem area, through the

skillful use of procedures capaHe of reliably accomplishing their purposes. Ex-

pertise in most fields of endeavor is typically associated with increases in the

repertoire of knowledge and techniques that make smooth, effortless responses to

high ground problems possible. As Bloom (1985) and others have pointed out,

mastery of such responses in complex areas of human functioning (e.g.. music.

sport, medicine, law) typically require at least a decade of devoted effort. We

assume that comparable effort is required before 'virtuoso' school-leadership

performonce on the high ground is likely. Of course, some of this preparation

usually does, and should take place before entry into a formal school-leadership

prsinon.The theoretical orientation that guided research on which this chapter is

based (as well as Chapter 6) is approximately similar to the orientation used

in many studies of problem-solving in other fields of endeavor. This is an

'information-processing orientation which explains how the individual mind

functions in response to a problem and why sonic peoples' responses seem more

expert than others. According to this orientation, individuals initiate the sort of

thinking Formally called problem-solving when they are motivated to accom-

plish a goal which they value like implementing a new program in their school

but have not achieved, To move toward that goal, they search their memory

for a pr,,cedore or routine that informs them about what to do. From this

perspective, the major resource available to problem solvers is their existing stock

of knowledge, skills and dispositions. In very general terms, it is the process of

searching through, selecting, reorganizing. reinterpreting, often adding to, and

then applying this existing stock of knowledge, skills and dispositions that we

call problem-solving.Studies of problem-solving, comparing novices and experts from fields otht r

than school leadership, have identified at least seven differences between them

Experts, as compared with novices:

ate better able to regulate their own problem-solving processes (Berliner,

196): this control appears to he what Schön (1983) refers to as 'reflecting-

in-action' and 'retlecting-on-action'


Dc veloping Expert Leadership jor Future Schools

possess more probh. m-relevant information (Berliner, 1986, Norris,1985) and have it stored in memory in a better organized, more richlylinked manner, thereby increasing its accessibility and extending itsapplication (Bereiter and Scardamalia, 1986)represent problems using more abstract categories (as opposed to moresuperficial features of the problem) with reference to more basic princi-ples (Berliner, 1986, Chi, Feltovich and Glaser, 1981, Voss et al., 1983):they also have better and faster pattern-recognition skills (Bereiter andScardamalia, 1986, Berliner, 1986)identify and possess more complex goals for problem-solving, and goalsrelated to action plans (Bereiter and Scardamalia, 1986, Berliner, 1986)spend more time at the beginning of problem-solving, planning theirinitial overall strategies; also, they are more flexible, opportunistic plan-ners during problem-solving, and are able to use a greater variety ofapproaches to a solution (Berliner, 1986, Norris, 1985)have developed 'automatic' responses to many recurring sequences ofproblem-solving activity (Norris, 1985)are more sensitive to the real requirements for solving the problem, andsocial contexts within which problems are to be solved (Berliner, 1986)

At the time the research reported m this book was undertaken, these differ-ences between experts and novices, which had been the product of studies inmany areas of activity, were of unknown relevance to school-leadershipproblem-solving. This deficiency gave rise to our research.

In order to describe how expert school-leaders solve problems on the highground, we draw on a series of related studies including both systematic litera-ture reviews and original empirical inquiries.= The three literature reviews havebeen summarized and updated in Chapter 2 of this volume and need not bedescribed in further detail here. Suffice to say that these reviews encompassed atotal of 135 studies, empirical and case-studies, conducted by other researchersbetween 1974 and 1988. which provided evidence about aspects of school leader-ship. Must of tl.e.se original studies provided information (from a variety ofperspectives) about either non-expert (typical) or expert (highly effective) school-leadership .r.iictices in dealing with problems on the high ground. These studies.however. do not recognize the distinction we consider extremely important, inthis book, between high ground and sw,unpy problems.

Our own empirical research relevant to this chapter included a .quence ofdiverse data-collection procedures aimed at answering five questions. The firstquestion was: What 'dimensions' or 'categories' would serve to organize descrip-tions of variations 111 the expertise of school-leaders' problem-solving on the highground? Intensive interviews with a random sample of two dozen elementary-school principals provided the data for answering this question. Given suchdimensions or categories (e.g., decision-makMg), the sewnd question askedabout the nature of expert, high ground problem-solving within each dimension.Opinion data from samples of principals, teachers. department heads and centraloffice staff were collected to assist in answering this question. These data wereintegrated with relevant findings from tile literature reviews by a working groupof researchers and practitioners from a variety of roles.

The third question of interest concerned alternative approaches to high


Expert School Leadership on the High Ground

ground problem-solving wh;ch might vary in their degree of expertness.Methods used to answer this question were similar to those used to answei theprevious question. Results were described as a 'profile of growth' (four levels) inschool-leaders' expertise on the high ground. Four separate studies, each usingquite different methods, were then conducted to verify the claim that thesealternative approaches to high ground problem-solving (our fourth question)differed in their effectiveness, as suggested by the profile. Interview techniqueswere then used to collect data from sixty principals in order to answer our finalquestion about similarities and differences in the high ground problem-solving ofelementary and secondary school-leaders.

The general pattern of results from this series of studies appears to be quiterobust. Other independent research efforts with similar intents have producedremarkably similar, although usually less detailed results (e.g., Hall et al., 1984,

Blumberg and Greenfield, 1980, Salley et a/., 1978, Dwyer et a/., 1984, Heck,I arsen and Marcoulides, 1990). Direct efforts by others to confirm claims fromour research have generally been supportive (e.g.. .1- rider, 1986).

Components of School-Leaders' Problem-Solving on the HighGround

Four components or dimensions of problem-solving appear to encompass a highproportion of differences in expertise among school-leaders on the high ground.These include:

Goals: the long term, internalized aspirations held by school-leaders whichform the basis for their decisions and actions; such goals might be derivedfrom a school-leader's visionFactors: those aspects of the school which are experiencLd directly bystudents, that significantly influence what the learn, and that can he in-fluenced by school-leadersStrategies: clusters of related actions or procedures used by school-leadersto influence factorsDecision-making: processe,, used to identify and choose from alternativegoals, factors, and strategies

These categories of high ground problem-solving processes emerged fromour own research data in a 'grounded' or quite natural way. That is, in the earlystages of our research we clustered e.pparently sinnlar types of information abouthow school-leaders solved high ground problems with these categories as the

resul:. After thc fact, however, there are good reasons why these categoriesought to have emerged a:; they did. Other research indicates that problems areincreasingly high ground in nature with the increased availability of significantamounts of knowledge relevant to the 'domain' or the problem area. On thesegrounds, it is reasonable that the 'Factors' component would turn out to accountfor considerable variation in school-leaders' expertise. This component encom-passes the school-leader's knowledge about curriculum, instruction, and otherrelevant aspects of the school's educational program, broadly conceived.

It is also evidcnt from litetature discussed previously that the availability of


Developing Expert Leadership for Future Schools

reliable procedures for solving poblems (transforming the 'givens' in the prob-lem situation to the 'goals') helps reduce the swampiness of a problem. Forthis reason, school-leaders' increasing mastery of the techniques described as'Strategies' allows them to demonstrate increasing degrees of expertise on thehigh ground. Finally, the 'Decision-making' component acknowledges that highground problem-solving demands less novelty and less 'mindfulness' thanproblem-solving in the swamp, as will be examined in Chapter 6. Indeed somefrequently encountered high ground problems may require no conscious thoughtat all. Others require choices to be made from known alternatives. But as long asthese known alternatives are a sufficient response, the problem remains on thehigh ground and decision-making captures the type of thinking required by theproblem. The 'Goals' component includes processes that are relevant to bothhigh ground and swampy problems.

Table 5.1 summarizes the results of our research describing four levels ofschool-leader expertise in solving high ground problems. These levels of exper-tise, described within the four components of our high ground problem-solvingmodel, have been given labels. The labels (systematic problem solver, level 4;program manager, level 3; humanitarian, level 2; building manager, level 1)

attempt to capture differences in the customary responses of school-leaders tohigh ground problems. Higher levels of expertise represent an accumulation ofskills, knowledge, and attitudes from lower levels on the part of the school-leader, as well as some significant shifts in the nature of beliefs. School-leaders athigher levels continue to engage in many processes evident at lower levels butsuch processes are usually parts of a more extensive repertoire, rather than thewhole repertoire.

Before describing the contents of what we refer to as the 'Principal Profile',several indirect but important findings should be noted. First, most school-leaders about whom data have been collected varied, to some degree, in thc levelof their expertise across the dimensions and subdimensions of the profile.Second, only a very small proportion of the school-leaders involved in ourstudies worked predominantly at the highest level described in the profile (about10 percent, based on data from about 200 principals across six school systems).Third, most school systems studied considered the lowest level in the profile todescribe minimally acceptable, rather than unacceptable, principal behavior.Filially, the profile focuses on processes that seem to be largely acquirable, givenadequate school-district support of the type discussed in Chapter 15.

Subsequent sections of this chapter describe variations in high groundproblem-solving expertise-1 comparing primarily the highest and lowest levels ofexpertise captured by our research. Much fuller descriptions are provided inLeithwood and Montgomery (1986) for elementary principals, and Lcithwood(19h6) for secondary principals. The relatively modest differences found betweenelementary and secondary school-leaders are left to the end of the chapter fordescription.

( ;oa Is

Virtually no conflict exists within current research, our own or others, regardingthe types of goals and goal-related processes associated with high ground exper-



Expert School Leadership on the High Ground

tise. Variations in expertise can be described in terms of the nature, sources, anduses of goals.

Nature of goalsHighly expert (Level 4) school-leaders had an implicit or explicit philosophy ofeducation including an image of what it means to be educated. This image wasconsistent with the values of the larger public served by the school and was likelyto encompass student knowledge and skill, as well as attitudes and values. Allcategories of outcomes were considered important by these school-leaders. Withthis comprehensive set of student outcomes as a frame of reference, experts' goalswere to provide the best education and best experiences possible for studentsserved by the school. For example:

For a K-6 (kindergarten to Grade 61 school, I would like to know thatwhen children leave us they are reliable, independent, productivelearners, and in terms of problems that are going to conic up. they can

solve problems.

Such experiences tended to extend beyond the formal instructional setting. Be-cause definitions of the educated person evolve with time, experts were know-ledgeable about changes relevant to goals for students, and receptive to changesthat might have helped to achieve such goals.

In contrast to experts, non-experts (Level 1) believed that teachers teach andthe principal runs the school. As one principal commented:

I have a really good staff. They know more about their subject matterthan me. So, basically, I stay out of their way and let them teach andspend my time making sure the school runs smoothly.

Maintaining a smooth-running ship was their main goal, bringing with it a

dominant concern for administrative logistics. While these principals sometimesjustified their focus on the grounds that students and teachers required a tranquilenvironment in which to work, running a smooth ship had become an end in itsown right. Change was a source of annoyance to these principals since it chal-lenged the maintenance of established rules and routines.

As the nature of school-leaders' goals grew in adequacy, their goals becameincreasingly based on a view of the educated person, increasingly consistent withthose of the larger school community, and increasingly open to change, in theface of reasonable evidence of the need to change.

Soura's of goalsSchool-leaders differed about the sources from which their goals were derived.As expertise increased, the sources from which their goals were selected becameincreasingly public in origin and greater in number. Experts systematicallyselected their goals from those espoused for students by agencies of the state(e.g., state education department or ministry of education), the local schoolhoard, and the perceived needs of the community and students served by theschool. Because the least expert principals valued running a smooth ship


Table 5 1 Expert school leaders problem solving on the high ground A summary of levels







Problem-Solving Components

Factors Strategies Decision-making

Se.ected from mu..,,o'epublic sources

1-1191,v amb,tiniis for al,students

Transformed :nto short-term goals for piann.ng

,lseci to actively increasecons,stency among sta'l

drections they..rst.ie

severasome of ....hit.-

are publicParticjlar focus on

excepflonal YucientsF -uuurages staff to use

goats 40r plann,ngConveys goals when

reuaested or as(,,, ar need anses

Attempts to .nfluence alifactors bearing on students

xpectations within factorsare specific

Expectations derived fromresearch and professionaliudgment

Attempts to ,nfluence 'actorsbearing or, the schoo;program

Expectations within factorsare specific

E xpectations are derived 4rompersonal and staffexperiences, andoccasionally from research

Uses a wde vahety ofstrategies

Criteria for choice,nclude goals, factors.context, and perceivedobstacles

Makes extensive use ofstrategies to achieve

Reiies on limited numberuf established, weil-tested strategies

Choice based on studentneeds (especiallyspecial students),desire to be fair andconsistent, concernto manage timeeffectively

Uses factor-specificstrategies that arederived largely frompersonal experience andsystem direction

Ski:led in use of multipleforms, matches forms tosetting and works towardhigh levels of participation

Decision processes orientedtoward goals of education,based on information frompersonal, professional andresearch sources

Anticipates, initiates, andmonitors decision processes

Skilled in use of severalforms selects form basedon urgency and desire toinvolve staff

Decision processes orientedtoward school's program basedon information from personaland professional sources

Anticipates most decisions andmonitors decision process




IL owlBuiidingManager

Derived fron belief inthe importance ofinterpersonal relationsto effective school =happy school

Goals may be ambitiousbut limited in focus

Goals not systematicallyused for planning

Conveys goals to others,f requested

Derived from personalneeds

Focus on schooladministration rather thanstudents

pursuit of instruction&goals considered to beresponsibility ofstaff, not principal

Conveys goals to others+ requested

Attempts to influence factorsbearing on interpersonalrelations

Expectations within factorsare ambitious butvague

Expectations are mostlyderived from personalexperiences and beliefs

Attempts to influencebearing on schoolappearance and day-to-day-operations (mostlynon-classroom factors)

Expectat:ons within factorsare vague

Expectations are derivedf'orn Personal expenences

Chooses strategies whichfocus on interpersonalrelationships

Choice based on view ofgood school environmentview of own responsi-bilities and desire tomake jobs of staffeasier

Makes little use ofsystematic factor-specific strategies

Chooses strategies basedon personal need tomaintain administrativecontrol and remain un-involved in classroom

Strategies mostly limitedto use of vestedauthority and assiststaff with routine tasks

Attends to factor-specificstrategies in asuperficial way, ifrequired to do so

Uses primarily participatoryforms of decision-makingbased on a strong motivationto involve staff so they willbe happy

Tends to be proactiveconcerning decisionsaffecting school climatebut largely reactive in allother areas unless requiredto act

Uses primarily autocratic formsof decision-making

Decision processes orientedtoward smooth schooladministration based on personalsources of information

Decision processes are re-active. inconsistent andrarely monitored




Developing Expert Leadership for Future Schools

(administratively), their goals wt.re derived from a sense of the administrativetasks requiring attention in order for this to be achieved. Goals did not oftenspring from curricular, instructional, or interpersonal considerations.

Uses of goalsInternalized goals serve as a potential focus for school-leaders in planning theiractions, and as a source of criteria for deciding what those actions will be.Research results suggested that as school-leaders increased in their expertise, therewas greater congruence between their espoused goals for school improvement,and their planning and decision-making. Less expert school-leaders sometimesespoused goals very similar to those of their highly expert colleagues, but seemedto ignore them in practice.

In addition to these personal uses for goals, highly expert school-leaderssought out opportunities to clarify goals with staff, students, parents, and otherrelevant members of the school community. They worked toward consensusabout these goals and actively encouraged their use in departmental and divisionalplanning. While those displaying non-expert forms of problem-solving some-times included such clarification of goals, it was common for these school-leadersto simply assume staff knowledge and agreement. One principal said simply,`I'm not very talky about these things'.


In order to accomplish their chosen goals, all school-leaders select for attentionaspects of the school - or factors - which, they believe, contribute directly tosuch goals. The task for the school-leader, then, is to determine what thccharacteristics of that factor need to be in order to help achieve their goal(s), andwork toward producing those characteristics. Our research identified eighteenfactors which were attended to by school-leaders. Of these, ten appeared to beardirectly on students' classroom experiences, largely through the teacher.

which teacher teaches which studentsthe objectives or outcomes teachers work toward with students, includ-ing the emphasis teachers place on different types of objectivesteaching strategies, including the types of learning activities thesestrategies are designed to provide for studentsthe types and amount of material and resources available, and the natureand degree of their usethe ways in which teachers assess, record, and report student perform-ance and experiencethe way in which time is allocated and the strategies teachers use to get(and keep) students focused on thc learning task, including studentdiscipline and controlthe subject matter, themes, or topics encountered by students in theirprogramsthe organization and appearance of the physical environment of theclassroom


Expert School Leadership on the High Ground

the nature of the relationships between students, and between the teacherand students in the classroomthe nature and degree of integration among curricular objectives within,and across, programs and grades

A second cluster of factors considered by expert school-leaders were thoughtto affect the experiences of students while in the school but outside the class-room:

the functions, assignments, and roles of people in the school and clas-sroom (including decisions about which teachers teach what grades andsubjects; the role of the psychologist, the janitor, etc.)the form and substance of communications and relationships with thecommunitythe nature and degree of organized, out-of-classroom experiences for thestudentsthe adult-role models provided by staff as individuals, and as theyintecacted with one another; the form and substance of communicationsamong staffthe form and substance of communications and relationships with out-of-school, school-system staffthe conduct of students while the school was responsible for themthe nature of the relationships teachers developed with students on theplayground, in the halls, and the like, and the role model provided byteachers in these relationships

Variations in principal expertise concerning factors were a function of thefactors principals selected for attention, and the source and nature of expectationsheld for these factors.

Factors of tnost concernAs school-leaders increased in expertise, the factors they attempted to influenceincreased in number and changed in focus. To a predominant concern for factorsbearing on school appearance, and the day-to-day operations of the school (Level1, building manager), especially outside the classroom (e.g., student behavior,material and physical resources) was added a concern for interpersonal factors(Level 2, the humanitarian). These, in turn, were subsumed, but not replaced, byattention to program-related factors, such as program objectives, and use of timeand its management (Level 3, the program manager). Those school-leaders withgreatest expertise (Level 4, systematic problem solver) paid attention to allfactors, eventually. This pattern of growth toward attention to all factors seemedto be directly related to, and perhaps explained by, school-leaders' goals. Themore closely linked to school improvement such goals became, the greater thelikelihood that factors selected for attention included those which would help

with school improvement.While experts tended systematically to address all factors, they did so over

an extended period of time. Short-term priorities often led to placing emphasistemporarily on a small set of factors. In contrast, non-expert problem-solving


Developing Fxpert Leaderchip for Future Schools

behavior (Level I) was characterized by long-term, consistent inattention tomany factors, and attention to others only when provoked by a crisis (e.g.,parental complaints about a curriculum topic).

Nature of expeLtationsAs school-leaders became more expert, their expectations within factors alsobecame more defensible and more consistent with prevailing professional judg-ment and the results of research. This suggests that such expectations, when met,stand a better chance of actually resulting in school improvement or goal achieve-ment. Expectations also became increasingly detailed or concrete with increasedprincipal effectiveness. Experts, for example, were better able to see whichspecial characteristics of their schools needed to be accounted for in formulatingexpectations they held for factors and, specifically, how such characteristicsmight influence those expectations in practice.

Non-experts (Level 1) had vague expectations regarding the limited numberof factors to which they attended. For example, concerning classroom manage-ment, one 'building manager' (Level 1) offered only:

I like to sec a classroom where there are a lot of meaningful things to do.

At Level 2, expectations tended to be high but still general; for example, staffwere expected to `cooperate with one another', but what such cooperatiorentailed was not made clear by the principal. Program managers (Level 3),although not concerned with the full array of factors, as were the most expertschool-leaders, were quite specific in their expectations for those factors ofconcern.

Sources ol expectationsInformation used in for ululating expectations also varied with expertise and camefrom many sources. Increased expertise, however, was associated with systematicrather than incidental or whimsical attention to non-personal sources. Expecta-tions at the least expert level varied according to what school-leaders believed tobe appropriate to the immediate situation. Such expectations were highly negoti-able and could be swayed by staff preferences, parental demands, administrativedemands, or the school-leader's interpretation of an educational trend. Withincreased expertise, knowledge of respected colleagues, and eventually research-based knowledge, were actively sought out and accommodated in formulatingexpectations. These sources of information undoubtedly increased the sophistica-tion or validity of principals' knowledge, hence the nature of their expectations.


Having identified factors associated with the achievement of valued goals,school-leaders still must act or intervene to influence selected factors in directionsthey consider most likely to assist in goal achievement. Our research showed thatschool-leaders employed a repertoire of both general-purpose and factor-specificstrategies to accomplish goals.

General-purpose strategies were considered by school-leaders as useful in



Expert School Leadership on the High Ground

influencing the characteristics cf almost any factor, depending very much oncircumstances in the school at the time action was taken. Such strategies estab-lished an appropriate background and climate within which more factor-qccificaction still had to be initiated to ensure goal achievement. Among the sevengeneral-purpose strategies we identified, four focused on keeping those involvedin decision-making willing to participate and well informed. These included:

the building and maintenance of interpersonal relationships and motivat-ing staffprovision of staff with knowledge and skillsfacilitation of within-school communicationfacilitation of communication between school Ind community

Two additional strategies, a general strategies that addressed the provision ofadequate organizational resources for staff work were:

allowance for non-teaching time for staffestablishment of procedures to handle routine matter.;

The final strategy, a general purpose strategy, was using vested authority; thepurpose for its use varied significantly between experts and non-experts.

After the appropriate background and climate were established, factor-specific strategies could begin to exercise a direct influence on selected factors.

They included:

program monitoringgoal setting, program planning and developmentprogram implementationstaff supervisionprovision of support resourcesdirect relationship with students

Criteria and emphasisDifferent levels of expertise among school-leaders were evident in the criteriathey used for choosing strategies. As strategic expertise increased, and goalsexpanded, the number and nature of strategies used over extended periods oftime also increased. This increase could be traced back to the changes in types ofgoals, from a focus on building management concerns through interpersonalrelations, to the school program, and finally to student achievement. Achievinggoals increasingly linked to student achievement eventually demanded attentionto all factors, as wc have already mentioned. Effectively influencing all factorsrequired the use of virtually all general-purpose and factor-specific strategies.Expertise also depended on school-leaders' ability to identify strategies thatwould impact on weak aspects of their schools' background or climate.

Non-experts (Level 1) on this dimension (strategy selection) of problem-solving needed to feel in control of administrative matters in their school. Suchcontrol was usually assumed through the use of vested authority. These school-leaders preferred not to be involved in decisions about curriculum or instruction,designating these as exclusively teachers' responsibilities. They also selected other


S u

Developing Expert Leadership for Future Schools

general-purpose strategies on the basis of intuitive judgments about what wasrequired to keep the school operating smoothly. For ezample, attention would begiven to interpersonal relationships among staff when a serious problem arose insuch relationships.

School-leaders at the next level of expertise (Level 2) sought out strategiesthat contributed to a warm, friendly climate in the school, often consideringpositive climate an end in its own right. They frequently gave considerableattention to such strategies, as well as being positive, cheerful and encouraging,accessible to staff, acting as a role model, and facilitating communication withinthe school, and between the school and community. When vested authority wasused, their reasons varied from a desire to make teachers' lives easier, by freeingthem from decision-making responsibility, to their convictions that some deci-sions were too specialized or important to be left to chance, such as schoolbudgets and teachers' record keeping.

A dominant concern for making fair, well-infbrmed, consistent decisionsand helping staff do the same was characteristic of program managers (Level 3).This concern motivated the systematic collection and distribution of informationrelevant to crucial decisions to staff. Such communication with the communitywas also viewed as an essential ingredient in building broader support for aschool's program.

The most expert school-leaders (Level 4) used a complex set of considera-tions in choosing their strategies, including the goals to be achieved, the factorsto be influenced, and characteristics of the people involved. They also consideredother activities already underway in the school, school and school-system norms,past experiences, and the nature of obstacles to be overcome. These concernswere used simultaneously and were viewed clearly as means rather than ends.Most general-purpose and task-specific strategies were used at some time byexperts to attain their goals.

Quality and skillSchool-leaders sometimes chose strategies well suited to factors in need of in-fluence and still failed to exercise much influence. One cause was the quality ofstrategies used. The effect of principals' actions was partially a function of thespecific procedures associated with their strategies. Principals increased in exper-tise, as their procedures became relatively more efficient in influencing factors(e.g., a single strategy influencing several factors) and as their strategies weremore readily used by others (many school-leader-initiated strategies depended onother members of staff to be completed). Strategies also were most effective asthey became more adaptable to changing school conditions. For example,program-planning procedures, useful across all areas of the curriculum, seemedto be generally more effective in stimulating subject-matter integration byteachers than strategies that were unique to subject areas.

Differences in the quality of strategies used is particularly evident in factor-specific strategies such as program implementation. Highly expert school-leadershad a strategy for program implementation that included well-refined, detailedsteps, applicable to many programs. Less expert principals either did not dealwith implementation (Level 1) or had no systematic approach to the process(Level 2).

It was still possible, however, for a school-leader to select a strategy poten-



Expert School Leadership on the High Ground

nally able to influence the factor(s) of concern, possess extensive knowledgeabout how to carry out the strategy, and still not obtain the desired effect. Thiswas the case when school-leaders' actual skill in use of the strategy was flawed ina crucial way. For example, some school-leaders knew that establishing goodrelationships with the community required listening carefully to parental con-cerns and patiently moving from such concerns, however expressed, toward afocus on how thcy were addressed in the school program. Yet these peopleallowed themselves to be frustrated with parental inquiries and frequently becamedefensive in their responses. Experts, on the other hand, were skilled in most ofthe general-purpose and factor-specific strategies. As school-leaders became high-ly skilled in their performance of a strategy, less conscious effort was required ofthem. This reduced the time required for them to respond to matters demandingtheir immediate attention (e.g., a report of drug use in the boys' washroom), andallowed them to attend to other problems for which solutions were less wellknown (e.g., increasing collaborative curriculum planning across departments in

the school).


Decision-making is a process that permeates the other dimensions of principalproblem-solving and helps account for the quality of those processes.

Differences in the way school-leaders chose their directions, selected aspectsof the school for attention, and decided to act accounted for much of thedifference in their expertise. Results of this research focused on two aspects of thedecision-making process. One was the context within which specific decisionswere made: the forms and procedures used in decision-making, principals' atti-tude and stance toward the process, and the monitoring of decision-making. Thesecond aspect concerned components of decision-making: how decisions weredefined, what criteria were considered relevant, and the use of information.

Forms and procedures for decision-makingThe most cxpert school-leaders knew about and demonstrated use of a range ofdifferent forms of decision-making in their schools. ('What I do depends on thesituation and the people involvecr) Sometimes they made unilateral decisions,sometimes they delegated the responsibility to others. Frequently, there wasextensive participation in the processes, with choices determined through consen-sus or, occasionally, by majority vote. The least expert principals made manymore unilateral decisions. When staff participated in the process, choices wereusually based on majority vote. The same school-leaders appeared to give littleconscious thought to which form of decision-making to use. In contrast, thosewith most expertise appeared to arrive at a choice of form by consciouslyreviewing: staff preferences and abilities, existing decision-rnaking practices inthe school, the nature of the decision to be made, and experiences from past

decisions.While highly expert school-leaders were consciously eclectic in the forms of

decision-making they used, they nevertheless had strong preferences towarddecentralization and extensive staff participation. Unlike those who were leastexpert, they used many decision-making occasions as opportunities to foster

8 267

Developing Expert Leadership .for Future Schools

conditions conducive to extmsive participation (staff willingness, skill, and aclimate in which the motives of those participating in decisions were widelytrusted). For instance:

In terms of my teaching staff, I would like them to feel that cooperativeaction is the best way to solve a problem, that no problem is insur-mountable, and that basically this is a pretty good place to be. Theyenjoy working here, so they want to go that extra mile.

These school-leaders were knowledgeable about how decisions were made indepartments or divisions in their schools and worked toward compatibility indecision-making processes at all levels in the school. Non-expercs tended to beout of touch with decision-makMg processes in which they were not directlyinvolved. Considerable diversity ir. such processes was typical within theirschools.

Variation in the procedures which school-leaders established for decision-making was also evident in our research. Lack of consistency in such procedureswas common among those with least expertise in decision-making. For example,sometimes their procedures allowed for different points of view to be heard,sometimes not; sometimes criteria for decision-making were made explicit,sometimes not By comparison, the most expert established procedures to helpensurt: consistent attention to: alternative points of view (including competingvalues), al; criteria relevant to the decision, clarification of the decision, andcollection of relevant information. Experts also had procedures (such as thedevelopment of a calendar, listing all major decision points in the year) foranticipating decisions and ensuring that needed decisions did not 'fall through thecracks'.

Attitude and stanie toward decision-makingLevels of expertise in this component of decision-making vaned in the extent towhich school-leaders sought out decision-making opportunities or reacted to thenecessity for decisions to be made. Experts tended to seek out decisions; theyviewed even minor decisions as opportunities to move incrementally towardtheir goals. They seemed able to anticipate a large proportion of decisions thathad to be made and used them to their advantage. Non-experts seemed unable orunwilling to forecast many upconung decisions. As a result, they found them-selves continually reacting to decision-making situations within a time-frameestablished by others. They rarely had enough time to make decisions carefullyand, nom surprisingly, tended to have negative attitudes toward change. Theirstance toward decision-making could be called 'crisis management'.

Monitoring decision-making. Those with least expertise in monitoring the process ofdecision-making and its consequences relied on their feelings (i.e., informalobservations, number of problems arising) about 'how well things were going'.Reactions to problems were in a piecemeal fashion with little effort to preventthem from recurring. At Level 2, staff satisfaction with decisions was frequentlyassessed. At Level 3, routine checks were typically made of school decision-making; special attention was given to how well the process met the principals'standards of fairness and consistency, and to the principals' perceptions of how



Expert Sthool Leadership on the High Ground

well school needs were being met. in monitoring, the most expert principalssystematically reviewed and refined the forms and procedures used. Informationwas usually sought out regarding the satisfaction of most of those affected by thedecision, including school staff; typically, resources or costs of the decisionprocess (e.g., amounts of time spent by department staff in selecting a newtextbook) and the contributions of decisions toward school goals (e.g., if thetextbook selected seemed to be the best one to contribute to program objectives)were examined.

Defining decision3 and seleaing criteria Variation in how school-leaders defineddecisions and the criteria they used appeared to be closely related to variations intheir goals. When running the school smoothly was the overriding concern, as inLevel I. school-leaders tended to take the path of least resistance in their decision-making: they responded swiftly to symptoms (e.g., placating a parent concernedabout the amount of homework given to his child) but ignored underlying causes(e.g., absence of a school-homework policy). The transformation of primarilymanagerial goals into criteria for dccision-making sometimes led to questionableemphases in the school. For example, some school-leaders responded to broadpressures regarding the basics, in such a way as to entirely ignore other equallyimportant goals. In some decisions where these school-leaders' usual criteriacould not be applied, choices were made intuitively with the claim that muchabout education was intuitive.

An overriding concern for a broad range of student outcomes, as in Level 4,was associated with efforts to uncover and clarify the fundamental causes ofproblems. Criteria, directly b:sed on the goals of education, included: the needfor individual programs, students' stages of development, and the need to balanceemphasis among knowledge, skill, and effective objectives. Other staff wereactively encouraged to use similarly oriented criteria in their decision-making.

Growth was also evident in how realistic and solvable the decisions definedby the school-leader were. Less expert school-leaders had a greater tendency toportray problem solutions as inaccessible (e.g., not enough time or money, nottheir problem, age of staff). The same basic problems were frequently cast inmuch more accessible tet ms by those with more expertise (e.g., weighing schoolpriorities, staff motivation and interest).

Use of information Those with least expertise in their use of information indecision-making collected little information within the school, except what wasrequested by district administrators. They tended to read report cards and wereopen to receiving other information but did not seek it. In contrast, expertsaccumulated information about most major functions of the school in a systema-tic way. They had procedures for routinely ensuring adequate information as abasis for major decisions. Further, they encouraged staff to do the same andexpected them to be able to identify the sources of information for their deci-sions. Those with least expertise only prcssed staff for information sources if thedecision was of special interest to them.

Information used most frequently by non-experts concerned building man-agement matters, and their responsibilities in such matters. This information wasusually available in the form of memos and policies from the district or ministryof education. Level 2 forms of problem-solving were characterized by seeking

8 A


Developing Expert Leadership for Future Schools

out information from staff, rrticularly about such issues as student morale andrelationships with parents. Frequent informal visits to classrooms was a typicalmethod of collecting this information. At Level 3, information was also soughtabout curriculum development and implementation activities in the school, andprogram requirements as outlined by the board or ministry of education. School-leaders at this level gathered this information through classroom visits, analysesof test results, reading of report cards, parental surveys, teacher plans, and otherformal assessments of student needs.

In their use of information, experts added to the processes described in Level3 a general knowledge about curriculum and education, gleaned from readingrecent research. This information was interwoven with school-specific informa-tion during decision-making. Experts also encouraged their staff to be familiarwith, and take account of, research-based information in their own decision-making; they attempted to keep staff well informed, by, for example, developinga handbook of procedures for school routines and carefully orienting new staff toschool expectations and procedures.

Elementary and Secondary School-Leaders:Differences on the High Ground

Our concern for the development of school-leaders stimulated us to inquire aboutpossible differences that might exist in the context of high ground problem-solving for elementary, as compared with secondary, school-leaders. If there aresignificant differences, these should be addressed, for example, in formal prepara-tion programs which, at present, pay very little systematic attention to schoollevel. Furthermore, only a handful of studies could be located which explicitlycompare elementary and secondary school leadership (e.g., Caldwell and Lutz,1978, Eastabrook and Fullan, 1978, Gersten, Carnine and Green, 1982, Johnston,1983, Licata and Hack, 1980, Martin and Willower, 1981, Morris et al., 1986,Newberg and Glatthorn, n.d., Salley, McPherson and Baehr, 1978, Willower andKrnetz, 1982). The results of these studies are sufficiently tentative to make theman unreliable basis for either understanding or action.

Our study (Lcithwood, 1986) of possible differences included both opiniondata from a random sample of sixty elementary and secondary principals, as wellas a comparative content analysis of our elementary and secondary school-leaderprofiles (Leithwood and Montgomery, 1986, Leithwood, 1986). Results iden-tified eleven differences distributed across each component of our high groundproblem-solving model (Table 5.2). Of these differences, six appear to be at leastpartly a function of school size rather than school level. Two of these school-sizedifferences provide potential advantages to thc decision-making of secondary(larger school) principals:


Greater opportunitics to delegate responsibilities because of more staffwith administrative responsibilities (number 10)Improved opportunities for effective decision-making because of thespecial resources (people and other types of resources) available withinthe school (number 11)



Table 52. School-leaders problem-solving on the high ground: Elementary vs. secondary


Difference Perceived Reason for Difference

1 Goals

1 Factors

3 Stiategies

4 Decision making

Ill nature of goals. the wide array of destinations forwhich secondary schools prepare their studentsmake their mission more complicated

121 source of goals secondary-school principals havesignificantly less Influence on school directionthan have elementary principals

131 program although the secondary principals'impact on the program is similar, theirrelationship to the instructional program is lessdirect than that of the elementary principals

14, nature of student relations with others studentproblems are generally less severe in nature forprincipals in elementary schools

15) support resources budgetary concerns occupymuch less time for elementary than forsecondary-school principals

be routine procedures secondary principals spendmore time on paper work than do theirelementary colleagues

17) elementary principals spend less time on record-keeping activities than their secondary-levelcolleagues

181 school-community communications parents aremuch more directly involved and far moreinfluential in elementary school than they are insecondary schools

q thin-school communication effectivecommunication is easier in elementary schoolsthan it is in secondary school

110) forms of decision-making elementary principalsare able to delegate fewer responsibilities thansecondary principals

till use of information secondary-school principalsare more likely to have support networks, andresource bases immediately at hand withinschool to assist in decision-making

need more diverse programnature of links with post-secondary ir, muttonsnature of policies to be implementednumber of secondary staffcomplexity of organizationscope and multiplicity of program and interests

rote of headsdegree of specialization number of staff

age of studentsmaller numbers of elementary schools allow forgreater closeness to students

vastly smaller sums of moneyelementary-school finances much less complex

timetabling activitiesgrading responsibilities

at elementary age, parents are more involved in ailaspects of their children's liveselementary schools generally are closer to and moreaccessible to the communitysecondary principals have more 'parts' to monitor,coordinate and meld together, and theiraccountability for larger sums of money necessitatestheir keeping volumes of Petalled statistical recordsfewer peoplesmaller plant

may be the only on-site administrator

various specialists are on site

Developing Expert Leadership for Future Schools

The remaining differences attributable to school size appear to offer potentialadvantages to elementary (smaller school) principals:

It is easier to communicate effectively with and among staff members(number 9)One's impact on program decisions and school directions can be moredirect in a smaller school where there are no other positions, such asdepartment heads, through which one normally works (m.. -nbers 3 and2)There is less time and effort required for record-keeping and other paperwork; timetabling is simpler, grading responsibilities are less demanding(nombers 7 and 6)Budgets require less effort to develop and monitor (number 5)

Two of the eleven differences in Table 5.2 (number 4 and 8) can be traced toperceived differences in the clients of elementary and secondary schools. Manysecondary school-leaders, in particular, reported spending large proportions oftheir time on student problems more complex and severe than those they zntici-pated were normally faced by their elementary-school counterparts (e.g., teenagepregnancies, higher incidence of substance abuse). These were perceived as prob-lems often extending beyond the competence or authority of the school-leader todeal with, although they were, at least initially, confronted with them. Parents,on the other hand, were viewed as having a more pervasive role in elementarythan secondary schools because of the more dependent natuure of youngerchildren and the closer identification of elementary schools with their immediatecommunity. This fact may add a dimension of complexity to the job ofelementary principals, compared to their secondary colleagues.

Finally, there appear to be differences in the mandate of elementary andsecondary schools that contribute to difference in the work of principals (number1). Elementary schools foster the general and relatively homogeneous develop-ment of young children and prepare them for entry into secondary-school pro-grams which can be known and controlled to a significant degree by the publicschool system itself. The secondary school, in contrast, is faced with the task ofincreasingly specializing the training it offers students, accompanied by an inevit-able press toward pragmatic and organizational compartmentalization: secondaryschools must offer highly diverse programs to students whose destinations in-clude the world of work and a plethora of post-secondary education institutions.

Summary and Implications for Leaders of Future Schools

This chapter has described variations in the processes used by school-leaders atfour different levels of expertise. to solve well-structured or high ground prob-lems. Four components or elements of such problem-solving, initially derivedfrom a series of studies (Leithwood and Montgomery, 1982, 1986, Leithwood,1986) were used to organize the description of these processes. These componentsincluded: the nature, sources, and uses of goals; classroom and school factorsdirectly influencing the achievement of goals; strategies used by school-leaders toinfluence factors; and the nature of decision-making about the other three corn-


S L)

Expert School Leadership on the High Ground

ponents. Within each of the problem-solving components, four levels of problem-solving expertise were described. The labels associated with each of these levelscapture the dominant orientation to problem-solving of school-leaders at thatlevel; building manager (least expert), humanitarian, program manager, andsystematic problem-solver. Differences were also identified in the context forhigh ground problem-solving of elementary as compared with secondary school-leaders. These differences appeared to be a function of either school size, natureof students, or the mandate:- for elementary and secondary schools.

What implication for leaders of future schools flow from this analysis ofschool leadership on the high ground? Three such implications seem to us to beespecially noteworthy: the value of a clear, 'working' vision of what the schoolshould be; the central importance of precise and defensible knowledge about theschool's main business of teaching; and the critical contribution of effectivecommunication in transacting the school's business.


While Chapter 3 was devoted entirely to the importance of a school-leader'svision, we extend several aspects of that discussion here, because the importanceof vision is especially clear in the context of high ground problem-solving. As wcasserted in Chapter 3, having a relatively clear, comprehensive picture of one'sschool in its future, ideal state is a powerful leadership tool. Indeed without sucha picture in mind it is difficult to imagine what would be the basic focus ofleadership in a school (or anywhere else, for that matter). Vision includes thegoals one aspires to accomplish for one's students, as well as those practices(factors) engaged in by school personnel when such goals are being accom-plished. A truly comprehensive vision might also include the strategies onewould use and the nature of the decision-making required to bring about thosepractices among school staffs and to accomplish desired goals for students.

While it seems essential for school-leaders to possess visions of their schoolsin their preferred states, such visions contribute little to school improvementunless they are widely shared. This is the case because a leader has only a limited,direct role in moving the school toward the vision. In tcrms of broad strategy,one may start with a vision and then convince others to accept it, start with avision and negotiate a version of it with others, or build a vision collaborativelywith relevant others (e.g., students, staff, district personnel, and communitymembers). Each of these broad strategies for arriving at a shared vision might bejustified as most suitable in some contexts. But the collaborative strategy is to bepreferred, in principle, since it offers the greatest promise for the sort of authen-tic, meaningful internalization of direction which actually guides people's dailydecisions and actions. Because internalized visions are such powerful guides topeople's decisions and actions, the time spent in their development contributes toclarifying the nature of the changes to which a school may aspire and to im-plementing those changes as well. In brief, it can be extremely efficient to 'wastea lot of time' building a shared vision. This is especially the case as demandslikely to be placed on future schools by increasingly diverse school clientsbecome more complex and uncertain. Or, as the president of a large successfulcorporation, quoted by Peters and Austin (1985, p. 285) put it: 'I think we'd all


Developing Expert Leadership for Future Schools

be better off if we spent more time articulating our corporate plans [vision) andless time on perfecting them'.

Tedinkal Knowledge

The 'technical core' of schools is their curriculum and its instructional delivery.Few issues have been more persistently debated than the proper relationshipbetween school-leaders and knowledge of the technical core. One position on thisrelationship is exemplified in the 'building manager' approach to high groundproblem-solving. Essentially it is an unconsumated relationship; the technologyfor teaching language arts (or any ther curriculum area) is virgin territory for thebuilding manager. That technology is married to the teacher and the buildingmanager wouldn't think of coming between thcm (even though he or she mayretain some 'carnal' knowledge from an earlier life as a teacher).

In contrast to the building manager's position, the current 'zeitgeist' forschool leadership is captured in the term 'instructional leadership' (e.g., Smithand Andrews, 1989, Duke, 1987). The term exemplifies what we believe to bethe appropriate orientation of future school-leaders toward knowledge about thetechnical core: that is, an intimate, long term, continually evolving and deepen-ing relationship. Why? Because, in the arcane language of cognitive sc ientists,'domain-specific knowledge' is one of the main pillars of expertise on the highground. School-leaders cannot just manage the process of problem-solving intheir schools and expect to accomplish the best for their students. To resort tothis approach implies an unrealistic expectation of teachers even under workingconditions quite different from those that currently prevail. This approach im-plies that teachers have the opportunity and motivation to seek out continuallythe best technical knowledge and to implement it. It implies, further, that theywill be able to introduce such knowledge at appropriate timcs during delibera-tions among staff members and make the case, in convincing terms, so as to betaken seriously; all without the substantive understanding and consequent in-tervention of the school-leader. This does not seem too likely (besides, no oneearning the salary of a school principal should get away with being of such littleuse). We consider a willingness and ability to work continuously on masteringthe technical core, to be a minimum requirem nt for leaders of future schools.We also consider this to be a realistic expectation because, in spite of thewarehouses full of writing about the technical core, there is not that much toknow! This does not mean that it would be easy to become a skilled performer inall areas of the technical core. It does mean that it is feasible to become an expertcritic and coach in many areas of the technical core.


Peters and Austin have also said:

7 4

Make no mistake about it. 'Techniques' don't produce quality products,educate children or pick up the garbage on time: people do, people whocare, people who are treated as creatively contributing adults. (1985,p. 199)



Expert School Leadership on the High Ground

Having just spent several hundred words on the importance of technical know-ledge for future school-leaders, we hasten to point out (more briefly) what seemsless obvious in practice than one would expect. The 'potential' of the leader'sknowledge is likely to be very disappointing in the face of poor-quality com-munication and unproductive interpersonal relations among members of theschool. This is evident, for example, in the amount of attention we had to devote

to interpersonal and communication strategies in our description of that compo-nent of high ground problem-solving. It is also evident in data suggesting thatschool-leaders spend about three-quarters of their time involved in communica-tions of some sort (Martin and Willower, 1981).

The skills of communication and of establishing productive interpersonalrelations are what Miles (1988) asserts to be the basic building blocks for schoolimprovement. These skills have to be reliably and continuously delivered or theschool-improvement effort breaks down. Future school-leaders will have tointeract productively with a wider array of clients and publics than is presentlythe case. And they will have to do so with less certainty than at present about thevalues, expectations, and customs of those with whom they interact. This willplace a premium on the possession of effective communication and interpersonalskills as vehicles through which their technical knowledge can be put to gooduse.


Are leaders made, or are they born that way? Th:s chapter has described what weconsider to be a large proportion of school leadership: expertise in solving highground problems. This characterization of such expertise makes our answer tothe perennial question quite clear: they are 'made', at least to a significant degree.One does not come 'naturally' to a productive vision and set of goals for one'sschool Nor does one have intuitive mastery of the technical knowledge of whichthis chapter spoke. To say that school-leaders are made, in this context, then, isto claim that there is much to be developed. One's expertise as a school-leader

ought to be considered a highly elastic phenomenon. Even two school-leaderswith comparable expertise will vary in the smoothness of their execution. Jack,the first systematic problem solver (Level 4) we identified in our research wasonly two years from retirement. He was so 'laid back' about what he was doingthat you had to watch very closely, to realize just how pervasive his presence wasand how much of an impact he made on his school. We have since observed

many less experienced, systematic problem solvers going about their work. They

do many of the things Jack did, but it requires much more overt energy on their

part Jack remains for us, the 'Wayne Gretsky of school leadership'. But it tookhim twenty-five years to look so slow, skating so fast.


I See Leithwood and Montgomery (1982), Leithwood (1982), Leithwood and Montgom-ery (1986, Chapter 1 I) and Chapter 2 this volume.

2 See Leithwood and Montgomery (1986, Chapter 13), Leithwood and Montgomery(1985), Leithwood (1986).

3 The following dcscription is adapted from Leithv. ood and Montgomery (1985).



Chapter 6

Expert School Leadership in theSwamp

Expert leadership in the swamp does not depend on the possession of an exten-sive stock of knowledge abcut the specific content of one's problems. For thisreason, the 'Factors' dimension of thc high ground problem-solving model de-scribed in Chapter 5 is not a promising element to include in a model describingexpert problem-solving in the swamp. Nor does expert leadership, in response toill-structured problems, spring from the more or less direct and skilled applica-tion of some set of tried and true procedures or techniques. Hence, the'Strategies' dimension of the high ground problem-solving model is of muchreduced value in explaining expertise in thc swamp. It is not that what wasincorporated within thc 'Factors and Strategies' dimensions, described in Chapter5, is suddenly irrelevant. It is the case, rather, that these dimensions do notcapture enough of what is important to account for much of the variation inschool-leaders' expertise in the swamp.

The model of problem-solving in the swamp, used as thc basis for describ-ing expert school leadership in this chapter, gives most weight to a set of generalcognitive strategies or thought processes (referred to as 'fluid ability' in contem-porauy theories of intelligence, Lohman, 1990). Also, the model includes ageneral disposition toward problem-solving. School-leaders' values and princi-ples are also critical parts of this model. Indeed, they are so pervasive that wetouch on them very lightly in this chapter, choosing instead to devote all ofChapter 7 to their nature and role in solving swampy problems.

In order to describe how expert school-leaders solve swampy problems bythemselves and in groups, we draw most directly on two studies. Both studieswere premised on the assumption that comparing and contrasting the problem-solving processes of expert school-leaders with their more typical or non-expertpeers adds precision to judgments about what is especially crucial in the experts'problem-solving processes. Accordingly, in both studies, the total sample ofprincipals was identified in advance as being either 'expert' or 'non-expert', usinga combination of techniques including nominations by senior administratorsfamiliar with their work, and interview techniques refined through the researchdescribed in Chapter 5. As it turned out, almost all principals labelled as expertsin both studies were 'systematic problem solvers', as designated in Chapter 5'non-expert' principals were 'humanitarian' in their orientation.

Study One, used to describe school-leaders' individual problem-solving



Expert School Leadership in the Swamp

(Leithwood and Stager, 1989) included a total sample of twenty-two principalswho responded to simulated problems which each of them identified in advanceas swampy. Their responses were tape recorded, transcribed, coded and contentanalyzed. Study Two (Leithwood and Steinbach, 1991a) is used as the basis forillustrating how expert and typical school-leaders solve problems in groups.Using a sample of ninc principals, data for this study were collected in thecontext of staff meetings which were tape recorded. Principals were asked todescribe what they were thinking about at various points during the meeting asthey listened to the tape immediately after the meeting (a technique called'stimulated recalf).

Results of these studies are described in the next two sections of this chapterorganized around the components of a model of problem-solving which emergedinitially from the data collected in the Leithwood and Stager (1989) study. Thesecomponents include:

Interpretation: principals' understanding of the specific nature of theproblem, often in situations whcre multiple problems may be identifiedGoals: the relatively immediate purposes that the principals are attemp-ting to achieve, in response to their interpretation of the problemPrinciples/Values: the relatively long-term purposes, operating princi-ples, fundamental laws, doctrines, values, and assumptions guiding theprincipals' thinkingConstraints: 'barriers or obstacles' which must be overcome, if anacceptable solution to the problem is to be foundSolution Processes: what the principals do to solve a problem (in lightof their interpretation of the problem, principles, goals to be achieved,and constraints to be accommodated)Affect: the feelings, mood and sense of self-confidence the principalsexperience when involved in problem-solving

These components ought to be considered more like categories of things school-leaders think about in solving a problem than steps in the process. Indeed, theprocess is quite fluid and components may be revisited several times; for exam-ple, efforts to develop a solution may reveal information that causes one tochange one's initial interpretation of the problem.

Flaying said the process is fluid, we hasten to add that there are limits to thefluidity. For example, initial interpretations of a problem sometimes lead tosolution processes from which one cannot recover. This would be the case if aproblem involving a teacher was interpreted as incompetence; such an interpreta-tion leads to goals and solution processes of a quasi-legal nature. The initiation ofsuch processes closes off the option of subsequently interpreting the problem asone of teacher development. It is important to 'get the problem as right aspossible' early in the process.


Developing Expert Leadership for Future Schools

How Expert School-Leaders Solve Swampy Problems byThemselves


Two aspects of interpretation seemed important from our research in Study One:the ways that principals attempted to understand problems, and principals' use ofanecdotes to interpret problems or to explain their understanding of problems.

Principals attempted to understand swampy problems in three ways: byrelying on past experience (as with the most structured problems); by collectingnew information; and/or by making assumptions. In Study One, the onlyassumptions made by experts concerned the hypothetical nature of the problemspresented to them and the experts were very explicit about these assumptions.On the other hand, non-experts tended to make assumptions rather than collectinformation and were somewhat less explicit about. their assumptions. This isillustrated, for example, in the analysis of one non-expert response to a probleminvolving a new principal taking over a school in which the existing principalwas very popular and the community was upset by her departure (described inChapter 4). The response revealed six assumptions, though they were not iden-tified explicitly as such: the previous principal was 'doing a good job'; theestablished procedures in the school 'were good'; changing procedures wouldlead to confrontations with staff; there were cliques to be dealt with 'ruthlessly%there would be a core of staff opposed to the principal's initiatives; and therewere 'people who control the school'. As this example also indicatcs, manyassumptions of non-experts were concerned with constraints or obstacles toproblem-solving.

These assumptions, along with an observed tendency to consider a numberof tangential or irrelevant issues, lcd to considerable 'floundering' on the princip-als' part in interpreting the least structured problems. For example, one non-expert, in considering the Principal-Entry problem, reported that he would spenda great deal of time and energy trying to find out why the previous principalhad been moved. To him, it was an issue of substantial importance whetherhis predecessor had been moved to another school or promoted to a super-intendency, and he mentioned the matte! repeatedly.

Principals frequently used anecdotes to explain and illustrate thcir approachesto problems. In contrast to non-experts, experts' anecdotes tended to be directlyrelevant to the problems at hand. Non-experts were more likely to recountdifficult experiences than highly successful ones.

In short, experts differed from non-experts in their ability to arrive at a clear,comprehensive interpretation of a problem, one th-t would enable them to geton with the actual solution of the problem. Ext did not appear to becomeinvolved in irrelevant issues and did not become dysfunctionally preoccupiedwith the feelings of others associated with the problem.


Both the nature and number of goals school-leaders attempted to achieve indealing with a problem distinguished experts from non-experts. Principals indi-



Expert School Leadership in the Swamp

cated, through their responses to the three least structurcd problems used inStudy One, that they were attempting to accomplish five types of goals. Thesegoals concerned staff, students and program, parents and the community,perceptions of others about the principals' expertise, and finding an appropriatebalance among goals.

Most principals had staff-oriented goals for problem-solving. Non-expertsmentioned this type of goal more often than other types. There weir certaindifferences between experts and non-experts in the particular staff-related goalsthat they cxpressed. For instance, non-experts were directly concerned with stafffeelings; in dealing with the Principal-Entry problem, one principal's goal was tohave 'good feelings about what you are doing ... and everybody happy withthat'. An expert, dealing with the same problem, believed that a more importantgoal was to have the staff understand his philosophy of education: 'I spent two orthree hours ... sharing [with staffl what I believed to be important abouteducation so that they would gain an insight into my philosophy'.

All principals identified goals related to students and programs in response toa problem about implementing a primary language program. They wanted toimprove the quality of student programs. With problems which had less obviousimpact on students, however, an important difference between expert and non-expert principals was apparent; only the experts outlined how they would includethe student and program categories of goals in their problem-solving. In aproblem about school consolidation, an expert identified implications for thegrouping of new students in terms of program, while the closest a non-expertcame to a student-oriented goal was 'selling the school' to prospective students.In the Principal-Entry problem, one expert was particularly emphatic about theprogram and student goals that should be addressed in solving the problem:

The whole fine reputation of the school ... and the confidence in theschool and the programs that arc in place arc at stake ... so you aredealing with the continuation of the program with the kids ... and soyou want that to be continued.

In this problem, also, goals of two experts included a concern with achieving abalance between their own and others' ideas concerning the nature of programs.

Most principals had goals for problem-solving that were related to parentsor thc larger school community. In this category of goals, experts were con-cerned with providing parents with knowledge, the better to understand theproblem and its eventual solutions. Examples of such goals from the school-consolidation problem were statements such as: 'parents get to know where theirkids are going to be', 'provided a vehicle for parents of new students to askquestions', and 'have parents questions answered'. Non-experts, on the otherhand, wanted parents to be happy and comfortable with the solutions, as indi-cated in such statements as: `to make people feel better about the new school' and'makc them feel comfortable'.

Two other goals were mentioned much less frequently and, for the mostpart. in connection with the Principal-Entry problem. One of these was to beknowledgeable about educational matters and the school. and to be seen by thestaff and community in this light: 'I would [want people to get thel impres-sion very, very quickly that I knew --hat the situation was, that I'd done my


9 5

Developing Expert Leadership for Future Schools

homework, and that I kne. the community and the program. I'd endeavor to bereally well informed'.

The second goal was to achieve an appropriate balance, in this case betweencontinuity (of the most desirable features of the present school) and change (ofthose features that could be improved); experts werc very concerned aboutdeveloping personal initiatives that would contribute to their new school, butthey wanted to do so in an appropriate manner.

In sum, principals' goal-related thinking suggested that experts pursued abroader range of goals and were more concerned than non-experts with know-ledge (their own and others) as distinct from feelings. They were also better ableto see the implications, for students and for program quality, of problems notobviously or directly concerned with students or programs, and were in generalmore concerned about achieving a balance among variou goals.

Principles and Values

This component is, in our view, an especially critical and pervasive aspect ofschool-leaders' problem-solving. Study One found this to be the case especiallywith experts. They were able to be much more explicit about the principles andvalues they used m their work. With such explicitness, principles and valuesserved as substitutes for knowledge in solving swampy problems.

As mentioned earlier, we save more extensive treatment of this problem-solving component to the next chapter.


Although the number of responses coded as constraints in Study One was rathersmall, the observed variation in then function was extremely important. Thereappeared to be little difference in the nature and operation of constraints for themost structured and least structured problems, but there were marked differencesbetween experts and non-experts with both types of problems. Experts did notindicate any constraints that would bring them to a standstill, whereas manynon-experts did. This was particularly so with least structured problems. Mattersthat non-experts indicated as constraints or obstacles were viewed by expertssimply as matters to take into account during problem-solving; potential con-straints were addressed through the solutions that thcy generated. Rather thanviewing public opposition as a constraint in the school-consolidation problem,for example. one expert simply noted, as part of the solution process:

I'd provide a vehicle for parents of new students to ask questions ...perhaps through letters and return slips of paper ... They have to havethe opportunity to get to know where their kids are going to go ...would have open-house public meetings, invite families and perhapsstudents to conic: and see the building, and Ihold I a meet-the-teachernight with some kind of general assembly.

In contrast, a non-expert dealing with the same question mentioned a num-ber of constraints, which are representative of those mentioned by others, such


9 6

Expert School Leadership in the Swamp

as a lack of knowledge or information ('so there's all sorts of questions that arecoming up that will need answers before I can really sit down and prepare');opposition from others (I'm going to be dumped on with a lot of this stuff',`people in the other school are probably going to fight it and be very anti thismove'); and a potential lack of resources ('it may be too much of a drain on theresources of the school, and we may not be able to handle all they think we can').

In sum, expert school-leaders assume that all swampy problems includeconstraints but that such constraints can be overcome as part of their problem-solving; they are not deterred or dismayed by the prospect of having to tacklesuch constraints.

Solution Processes

Experts indicated, in Study One, marked efforts at thinking through solutionprocesses in considerable detail and planning how the process would be carriedout. In response to the Principal-Entry problem, one expert stated:

But the actual steps ... I guess is what wasn't really clear to me exactlywhat I would do ... I would have to sit down and I'd have to think itthrough and plan it out and look at the individual steps and where Icould go with each one.

Non-experts showed very little evidence of attention to planning. Mostprincipals met in some way with staff, parents and/or students, depending on thenature of the problem. All but a few non-experts stressed the importance theyattached to the details of th.. solution process in the context of such meetings.Again, in response to the Principal-Entry problem, one expert noted, forcx a nip le:

In my first communication with parents, I would talk about a lot of thegood things that the former principal had started and what my plans are.I would want to state my particular goals over the next school year.

Experts and non-experts differed substantially, however, in their orientationto such meetings a matter explored in greater depth in the next section of thischapter. Experts provided evidence of high levels of consultation in working outa solution process. One of the most marked features of the Principal-Entryproblem was that two of the three experts who solved it described explicitly (andthe third expert alluded to) very extensive consultation about their new schoolwith the best possible informant, the outgoing principal; none of the non-expertsdid this. Experts also stressed a broad array of features of the problem to beexamined through such consultation. One expert, for example, in solving theschool-consolidation problem, mentioned numbers of students at each gradelevel, nature of students, community support, projected enrollments, schoolorganization, staff ratio, student groupings, budget preparations, and plans fortextbooks and supplies. In general, the consultations of the experts in the leaststructured problems had to do with information collection. Non-experts con-sulted less frequently than did experts. When they did, some may have done so


9 "

Developing Expert Leadership for Future Schools

because they were confused as to how to proceed For instance, two non-expertssought the support of the superintendent before they even attempted to deal withthe problem, rather than going to him or her for a specific purpose after they hadbegun to formulate a solution.

All but one principal attached some importance to information collection asa solution activity. Experts awarded this activity greater importance and spentmore effort on it than did non-experts. Non-experts never followed up theconsequences of solution activities. Although it was not a marked aspect (as was,for example, information collection) of the problem-solving of experts, three ofthe six did include some sort of monitoring or follow-up process.

In sum. experts and non-experts differed markedly in their solution pro-cesses. Experts spent more effort planning for the solution process and identifiedmore detailed steps to be included in the process than did non-experts. Expertsalso consulted others more about the solution and attempted to elicit widespreadsupport for it. They stressed the value of careful information collection.


Finally, the mood or affective states of prindpals in approaching and solvingproblems was quite different. Experts were invariably calm and relatively con-fident in their own abilities. Non-experts, when responding to swampy prob-lems, were sometimes fearful, often not confident, and occasionally somewhatbelligerent and/or arrogant.

Responses of an expert and then a non-expert, collected in Study One,describing setting school objectives (a problem that both regarded as easy) illus-trate several of these differences.

From the expert:

This did not present a difficulty for me at all, because, first of all, I

believe in having staff give their input in terms of providing directionsfo, the school. So I had no difficulty with that, and I've done that as aprincipal from day one. I don't feel, just because I'm the principal, that Iset the entire goals of where the school is going. There are certainlymany, many directions that come from staff, and always have, in myexperience. And, as a teacher, I wanted input. And so that was just asimple situation for me, not even a problem.

And in contrast:


Setting school objectives, that's a process that's built in, it's a processthat I've done for a number of years. I can show you the outcomes, andI can show how ... You know, it's there, it's a book. I could give it toyou. it's done If I were coming to a new school. I believe as theprincipal, you say, 'Here's what I stand for. Here arc my expectations'.And you say all that stuff up front at the opening staff meeting. Any-body who waffles and says. `I'm going to spend six months lookingaround' is either lying or stupid.


Expert School Leadership in the Swamp

Table 6 1 summarizes experts' processes, as they have been described in thissection.

How Expert School-Leaders Solve Swampy Problems inCollaboration with Others

Corporate entrepreneurs - single minded individuals that they are - stillget their projects done by crafting coalitions and building teams ofdevoted employees who feel a heightened sense of joint involvement andcontribution to decisions. ... Masters of change are also masters of theuse of participation. (Kanter, 1983, p. 241)

The results presented in the previous section, describing principals solving prob-lems by themselves, showeu that completely autonomous problem-solving isrelatively rare among expert school-leaders. While some components of theprocess are carried out by the school-leader alone, this is not often the car.- for thesolution-process component, in particular. Indeed, our research suggests thatnovice school-leadcrs often believed that they were hired because they knew thcanswers and tended to solve problems without much involvement of othersbecause they thought that was what was expected of them. Increased experience,expertise, and leadership responsibilities, however, tend to be associated with adecrease in isolated problem-solving, at least for non-routine, non-technicalproblems (Leithwood and Stager, 1986, Leithwood and Steinbach, 1991a, Leith-wood and Steinbach, in press, b). As an expert secondary principal said to us:

There are very large decisions, and I'm trying to think of one - there arealmost none that would not be made by a group of people. Andnormally that group of people would be thc group that's interested inmaking a decision.

Some degree of collaboration with others, then, is the norm as expertschool-leaders solve swampy problems. But there is a significant difference in thedemands placed on the problem-solving of school-leaders by modest degrees ofcollaboration (e.g., asking teachers for information about a specific element of aproblem) as compared with fully participative procsses (e.g., involving teachersin interpreting the problem through to finding a solution). This is evident in theresults reported by Duke and Gansneder (1990) showing significant variationsacross individual teachers in their preferences for extent of participation and typeof problem they consider appropriate to participate in solving. How expertschool-leaders solve problems in this fully participative manner is the focus ofthis section.

Using components of the problem-solving model introduced in the previoussection, we describe how four expert and five non-expert principals in StudyTwo solved swampy problems (a real problem they were working on in theirschool at the time of the research, and different in each case) in collaboration withothers.



Developing Expert Leader3hip ;or Future Schools

Table 6 1 Expert school-leauurs problem-solving processes Without collaboration

Components Indicators of Effectiveness



Principles / values


Solution Processes


Classifies problems (in terms of urgency, personal time priority,and who should be involved)Indicates awareness of possibility of more than oneperspective (e.g., educational, political, legal, financial) on thenature of the problem and, where appropriate, shifts amongperspectivesIndicates awareness of goals and values of other individualsand organizations who are involved in the problemAccurately detects similar and different (from past situations)features of problem situationIs responsive to particular opportunities and constraints in asituation

Identifies goals from various sources that are appropriate tothe problemBalances a number of goals, including those of others, infinding solutions

Refers to principles or values in considering problemRefers to personal and professional principles or values inconsidering problemIndicates, in complex situations, which goals and principles areinviolable and which are flexible and negotiable

Indicates only those constraints which do in fact exist in thesituation (e.g., does not make erroneous assumptionsreprding constraints)A:curately identifies constraints in situationEi(her converts these constraints into subproblems forsolution, or interprets problem differently in order to avoidconstraints

Collects appropriate amount of information from optimalsourcesDevelops and/or monitors a deliberate solution planUses and, in some circumstances, shares a framework fortackling problemsInvolves others in problem-solving where appropriateIndicates in group problem-solving situations a concern forinsuring shared understanding of problem situation and goalsof all involved

Appears confident, calm, and 'centered'Enloys problem-solving as providingopportunities


Expert and non-expert principals differed significantly in their approach to prob-lem interpretation. For the most part, these are differences in degree. Evidencefrom the non-expert principals revealed no signs of conscious reflection on theirown problem-interpretation processes. While such reflection was not extensive


10 0

Expert School Leaders lup m the Swamp

on the part of expert prmcipals either, one explicitly talked about the importanceof having a clear interpretation of the problem

To me the critical part is identifying what the problem is ... and theproblem is different for different people.

Expert and non-expert principals did vary substantially on the extent towhich they took into account the interpretation others had of the problem theywere addressing. Two of the four experts explicitly checked their own assump-tions and actively sought out the interpretations of their staff members as well. Incontrast, none of the non-expert principals did this and three of the five assumedthat their staff had the same interpretation of the problems as they had.

Another difference in problem interpretation between expert and non-expertprincipals concerned the context in which problem interpretation took place.Most expert principals viewed the immediate problem they were addressing inthe context of thc larger mission and problems of their schools. For example, oneeffective principal noted in his introduction to the topic:

Discipline is a school thrust landI the PA day Iwe had on it] was toobrict Wc need to equip ourselves with more strategies, refine our skills,beciuse of the greater challenges we (nowj face in the classroom, forexample, the integration of special kids.

Filially, a particularly striking difference between expert and non-expertprincipals was the much greater degree of clarity experts had about their inter-pretation of the problem, and their ability to both describe their interpretation totheir staffs and to indicate the reasons they had for such an interpretation. Anexpert principal said, for example:

Earlier in the fall we had a discussion about a house system and some ofthe ideas were of some concern to one division more than another. I

don't think you were really able to come to a decision that it should beschool-wide and yet there was a feeling that it might very well work outthat way ... We took it to a lead teacher meeting ... with someguidelines I had. We've done sonic revising of those guidelines and whatI will give you is an outline of that and some of the ideas behind it.Hopefully, then the meeting will conic to you as a house system thatcould really work for Ithis school I. I favour for that to happen if it canunder the whole philosophy we have here of lots of participation andlow key amount of competition.

This does not really deal with school teams ... We're talking aboutin-school activities and its not just sports teams and I think that's part ofwhat S is after too. Each of you will get a sheet with a revised set ofguidelines and thoughts and S will carry forth from here.

Non-expert principals frequently were unclear about their interpretations and haddifficulty in explaining the reasons for the interpretations they held.


1 0 1

Developing Expert Leadership for Future Schools


Expert and non-expert school-leaders shared many similarities in this componentof their problem-solving. Both had multiple goals for problem-solving: usuallyfive or six goals in relation to any given problem. For example, in response to aproblem regarding results of a survey conducted about the school one principalvoiced these goals:

share the findings [of the survey conducted]identify one, two, or three areas that we see as targetsrelate these target areas to overall school goals that will be formulatedshortlyinvite staff input'I'm not looking for solutions' looking for agreement among the group

Furthermore, both groups of school-leaders made a point of sharing their owngoals with others involved in problem-solving. Eight of the nine principals alsoset goals not only for the problem to be solved but also for the meeting in whichcollaborative problem-solving was to occur. One non-expert principal estab-lished goals only for the problem and gave little or no attention to the process forproblem-solving.

However, there were two differences between expert and non-expert prin-cipals with respect to goals. One of these differences concerned the relationshipbetween the principals' goals for problem-solving and the goals that other staffmembers held. Expert principals indicated a strong concern for the developmentof goals that could be agreed on by both themselves and their staff. For example:

I want this to go, quite frankly I do, but I want it to be something thatthey have sct up the way they can make it work. It's not something Iintcnd as principal to give out great big awards for; it's something thatI want the teachers to feel comfortable that they arc able to build aprogram.

Non-expert principals, in contrast, werc concerned only with achieving theirown goals and with persuading their staffs to agree with them about what thosegoals should be. As onc principal said: 'I think I got them [the teachers] toidentify the several key areas that are relevant from my point of view.'

The second difference between expert and non-expert principals was thestake held by the principals in a preconceived solution. Non-expert principalswere often strongly committed to such a solution prior to entering the 'collabora-tive' problem-solving process and constantly manipulated the process in an effortto gain support for that solution. Expert principals, on the other hand, had muchless stake in any preconceived solution. They wanted the best possible solutiOnthe group could produce and took steps to ensure that such a solution was found.For example, one principal said:

I can either see it happen or let it wash through and say it was a goodtry. maybe another time. I'm willing to go with It other way.

Expert School Leadership in the Swamp

As already noted, we leave a detailed treatment of this component to Chapter 7Suffice to say at this point that experts made more use of their values than didnon-experts and relied on different values in some of their problem-solving thandid non-expert principals.


Expert as compared with non-expert principah were better able to anticipate theobstacles likely to arise during group problem-solving. Non-expert pri iIs

either did not anticipate obstacles or they identified relatively superficial , i-

des. Even when they did anticipa:e obstacles, non-expert principals rarely consi-dered, in advance, how they might respond to those obstacles should they arise.This is in contrast to expert principals who planned carefully in advance for howthey would address anticipated obstacles, should they arise. In addition, expertprincipals adapted and responded in a flexible way to unanticipated obstacleswhich arose. For instance, during the 'stimulated recall' sessions involved inStudy Two, one principal said:

The point he is making which I hadn't taken into account ... I'mbringing up discipline, so obviously it's interpreted that I'm not happywith discipline in the school.

The prthcipal went on to say during the meeting:

I am glad that point has been raised. By doing this. I'm not saying thatthings are falling apart. People arc on top of things and I appreciate that.At the same time, it's something we must continually be at.

He then provided a personal example of a difficulty he had in handling adiscipline problem.

Expert principals tended not to view obstacles as major impedimr.nts toproblem-solving in the same way that our previous evidence has suggestednon-experts do. Furthermore, whereas experts were concerned to learn and build

on the perception of their teachers, non-experts viewed differences in their

perceptions and those of their teachers as frustrating constraints. This is notsurprising given their preconceived ideas about what goals were to be achieved at

the meeting.

Solution Processes

The greatest differences between expert and non-expert principals were foundwithin this component of their problem-solving. Non-expert principals rarelyplanned for collaborative problem-solving. In contrast, experts developed careful

plans, as is evident in these remarks:


Developing Expert Leaders lup for Future Schools

[First] I'll share with them what the problem is for me [Then] I'llhave them working in small groups to identify the problem or whatthey see as the problem brainstorm some ideas of what they think theproblem is ... After we have a look at what the problems are we canlook at some potential solutions.

During the process of collaborative problem-solving, in he context ofa staffmeeting, experts ensured that the stage for effective collaborative problem-solving was set by providing a clear, detailed introduction to the problem and itsbackground. This was not a matter given much attention by non-experts, one ofwhom provided no background at all to his staff. Although one non-expert diddevelop a clear plan for problem-solving at the meeting, he did not share thatplan with his staff Most of the experts shared not only the background to theproblem but the process that they were suggesting for its solution how themeeting would be conducted, for example.

Part of the process for conducting the meeting, on the part of experts, wasthe careful checking with staffs regarding their interpretation of the problem, andthe extent to which their assumptions were shared by staffs. Non-experts invari-ably assumed that others had the same interpretation of the problem which theyheld and did not check to see whether or not that was the case. Non-experts,because of their prior commitment to a set of goals and a problem solution,tended to either argue overtly for their own view of what the solution should beduring the meeting with their staffs, or manipulated the meeting subtly so that itsupported such a view. For example:

I did cut off one person in the group who spoke only once, and tried tosolicit more time to pursue all the data but that was not part of myagenda.

This principal also deflected the need that teachers had to vent feelings of dis-appointment or outrage over prior results in a survey, even though he had theopportunity and need to do just that.

I went through that stage and now look at it as a reality ... Let's notstay with the raw numbers here; let's get past that.

Other non-experts changed topics or called on teachers who used the strategiesthe principal wanted accepted.

Expert school-leaders were able to make clear their own view of the prob-lem without intimidating or restraining their teacher colleagues. In addition, theywere open to new information and, if such information warranted it, wereprepared to change their views of what the solution should be. Experts, incontrast with non-experts, facilitated collaborative problem-solving by synthesiz-ing the views of others, summarizing progress in the meeting from time to time.providing clarification as needed, and gently prodding the group to keep on task.

None of the non-experts considered how to follow up on decisions madeduring thc meeting, whereas this was done by almost all experts.


Expert School Leadership in the Swamp


Both experts and non-experts usually appeared to be calm and confident duringthe pro;. -m-solving process with their staffs. One non-expert's frustration wasvisible on one occasion. There were substantial differences, however, in theamount of anxiety or frustration actually experienced (but not demonstrated) byprincipals. This was evident in their discussion following the staff meeting. Inthat context, the majority of non-experts expressed some frustration over themeeting. Their source of frustration was the unwillingness of the staff to agree onwhat the problem should be:

Inside me I say. your question is not relevant. It has nothing to do withwhat we're doing. Damn it, I don't need that .. . Outside I hope I didn'tshow it.

One non-expert also showed sonic: signs of insecurity about his own ability tosolve problems collaboratively.

I guess the problem that I had in the first place was to find a problem foryou to see. So I guess the kind of thing I want is some feedback (about]how I conduct meetings on all problems.

There were few signs of such frustration on the part of experts, although oneexpressed a mild form of frustration with her own seeming inability to help thestaff to arrive at a suitable solution. Experts and non-experts made use of humorto diffuse tension and to clarify information during collaborative problem-sol ving.

Table 6.2 summarizes the processes used by expert and typical school-leadersas they solved swampy problems in collaboration with their teacher colleagues.

Summat y and Implications for Leaders of Future Schools

This chapter has described the processes used by expert school-leaders to solveill-structured or swampy problems. Six components or elements of suchproblem-solving, initially derived from Leithwood and Stager's study (1989),were used to organize the description of these processes. These elements includedproblem interpretation, the setting of goals, the nature and use of principles andvalues, constraints, solution processes, and mood. Processes used by samples ofexpert school-leaders were contrasted with the processes used by their non-expertpeers, in order to highlight the most significant aspects of expertise. In thissection, we revisit four features associated with what has been learned aboutexpertise in the swamp that are especially noteworthy for leaders of futureschools.

Cognitive Flexibility

One noteworthy feature, evident in how experts solved problems both indi-vidually and in groups, is cognitive flexibility. By cognitive flexibility we mean,



Table 6 2 School-leaders' problem-solving processes. With collaboration

Components Expert School-Leaders Typical School-Leaders





understands importance of having a clearinterpretation of problemseeks out and takes into account theinterpretation others have of the problemimmediate problem usually viewed in its relationto the larger mission, and problems, of schoolhas a clear interpretation which they can describeto others and rationalize

has multiple goals for problem-solvingshares own goals with others involved inproblem-solvinghas goals for both the problem and the meetingin which collaborative problem-solving occursstrong concern for the development of goals boththe principal and staff can agree toless of a personal stake in any pre-conceivedsolution. wants the best possible solution thegroup can produce

see Chapter 7

accurately anticipates obstacles likely to ariseduring group problem-solvingplans in advance for how to address anticipatedobstacles should they ariseadapts and responds flexibly to unantic patedobstacles which arisedoes not view obstacles as malor impediments toproblem-solving

no conscious reflection on this matter

assumes others share same interpretation

tendency for problems to be viewed in isolation

less clarity about their interpretation, difficulty inexplaining it to others

has multiple goals for problem-solvingshares own goals with others involved inproblem-solvinghas goals for both the problem and the meetingin which collaborative problem-solving occursconcerned with achieving only own goals andgetting staff to agree to those goalsoften strongly committed to a pre-conceivedsolution and attempts to manipulate groupproblem-solving so as to result in support for thepre-conceived solution

see Chapter 7

does not anticipate obstacles or identifiesrelatively superficial obstacles

rarely considers in advance how to respond* .

those obstacles that are predicted I

Solution Processes


1C. td..

has well developed plan for collaborativeproblem-solving (meeting)provides clear, detailed introduction to problemand its background to collaboratorsoutlines clearly the process for problem-solving(e.g . how meeting will be conducted)carefully checks collaborators interpretations ofproblem, and own assumptionswithout intimidating or restraining others, clearlyindicates own view of the problem andrelationship with larger problemsremains open to new information and changesviews, if warrantedassists collaborative problem-solving bysynthesizing, summarizing and clarifying asneeded, and by keeping group (gently) on trackensures that follow-up is planned

always appears to be calm and confident

has hidden anxieties, usually the result of inabilityto find a workable solution

invariably treats others politely

uses humor to diffuse tension and to clarify,nforrnation

rarely plans for collaborative process and mayvalue 'spontaneity'does not provide clear introduction to problemwhich occasionally is missing altogetheris not likely to share plan for meeting withcollaborators if plan existsassumes others have same interpretations ofproblem: does not checkarguec slubbornly for own view or 'orchestrates'meeting so that it supports such a view

adheres to own view in the face of competingviewsshows limited action to assist collaboration andmay seriously underestimate time required forcollaborators to explore problemrarely considers plans of follow-up

usually appears calm but frustration mayoccasionally become visiblefrequently feels frustrated, especially byunwillingness of staff to agree with principal'sviewsoccasional signs of insecurity about own ability tosolve problemsuses humor to diffuse tension and to clarifyinformation



Developing Expert Leadership for Future Schools

for example, the ability to exe.cise control over one's own thought processes andfeelings, the ability to detect that one has gone down a blind alley and thewillingness to back up and consider other alternatives. Cognitive flexibility alsomeans being open to the views of others, not being held hostage by one's ownprevious experience (e.g., not seeing every new problem as just a version of anold problem) and being able to change one's interpretation of a problem whenconfronted with new information. Cognitive flexibility is also a function of one'sattitude toward problem-solving; one displays flexibility when `problems' areconsidered interesting 'challenges' and 'opportunities' to accomplish one's goals.Many of these aspects of cognitive flexibility were observed among experts in thetwo studies reviewed in this chapter. All were evident in another of our studiesinquiring about the meaning of cognitive flexibility in the context of schoolleadership (Leithwood and Stager, 1989). These aspects are consistent with thefindings of research on social cognition in other fields of activity (Showers andCantor, 1985).

Stager and Leithwood (1989) also found evidence among typical principalsof the kind of cognitive 'errors' associated with inflexible thinking in other fieldsof activity, as reported by Nisbett and Ross (1980) and others. Such principals,for example, were prone to set priorities for problem-solving based solely onhow emotionally vivid or immediately pressing the problems seemed to be (1keep prioritizing. When there arc no problems in the in-basket or walking intomy office, I go and play with the kindergarten's ...'). They also tended togeneralize from small or biased samples (e.g., involve only like-minded staff insetting school objectives), and failed to determine the actual cause of problems,or that some problems had important elements that were unique (`No brand-newproblems come to mind. I sometimes say "I was doing this exact same thingtwenty ycars ago".'). These are examples of inflexible thinking, of which therewere others, not evident in the problem-solving of expert school-leaders.

By way of summary, then, our studies of problem-solving in the swamppoint to the fundamental importance of cognitive flexibility, as one basis forexpert school leadership, now and in the future. Such flexibility involves a totalavoidance of cognitive errors and an ability, noted by many authors (e.g.,Morine-Dershimer, 1986, Nisbett and Ross, 1980, Schön, 1983) to make veryfine discriminations among details of particular situations. Second, cognitiveflexibility involves controlling one's negative moods and approaching problemsituations with an air of calm confidence. Klemp and McClelland (1986), instudying characteristics of intelligent functioning among managers, found thatthe 'competency' of self-confidence was absolutely essential and served to drivethc other intellectual competencies. Third, cognitive flexibility involves beingresponsive to the possibilities in the situation, affording a clear illustration ofwhat Sternberg and Wagner (1986) would regard as 'practical intelligence'. Theseauthors use Neisser's (1976) definition of 'intelligent performance in naturalsettings ... as responding appropriately in terms of one's long-range and short-range goals, given the actual facts of the situation as one discovers them' (p. 137).

It should be noted that cognitive flexibility never involves what Bolman andDeal (1984) refer to as 'overresponsiveness' or 'spinelessness'. Expert school-leaders have core values, beliefs, and goals (or vision) that arc inviolable. That is,they display extreme cognitive flexibility, but they are always guidcd by visionand a coherent set of values.

92 110

Expert School Leadership in the Swamp

Using the Capacity of the School Staff to Invent Powerful Solutions

The ability to make the most of the talent of those within the school itself inidentifying powerful solutions to swampy problems was evident in our researchon expert group problem-solving. 'Bounded rationality' offers a powerfultheoretical perspective from which to appreciate how collaborative problem-solving can lead to better solutions. The phrase, initially coined by Simon (1957)was intended to draw attention to the limitations in a person's capacity to processinformation in the face of the complex demands placed on that processing byfrequently encountered problems. The limited capacity of short-term memorywas of particular interest to Simon (1957) and others who elaborated the idea. AsShulrnan and Carey (1984) explain, however, bounded rationality focused exclu-sively on individual thinking and did not adequately recognize how individuals'participate in jointly produced social and cultural systems of meaning thattranscend individuals' (p. 503). Because human rationality 'whether bounded ornot, is practiced in the context of social exchange and human interaction' (p.515), a view of people as collectively rational is offered as a better conception ofproblem-solving in many life circumstances From such a view, problem solversuse others to compensate for their own limitations. They do this by transform-ing, redefining, and distributing parts of the problem task to others in theworking group, in an opportunistic way according to each individual's uniqueabilities. More specifically, under ideal collaborative problem-solving conditions,better solutions seem likely to be the result of, for example:

A broader range of perspectives from which to interpret the problem(Expert school-leaders do this when they actively seek out staff inter-pretations, are explicit about their own interpretation, and place prob-lems in a larger perspective)An expanded array of potential solutions from which to choose (Expertschool-leaders foster this by assisting group discussions of alternatives,by ensuring open discussion, and avoiding commitment to preconceivedsolutions)A richer, more concrete body of information about the context in whichthe problem must be solved (In our research, experts did this by activelylistening to staff views, clarifying, and summarizing information duringmeetings)The reduced likelihood of individually biased perspectives operating inthe solution process (Experts assisted with this by keeping groups ontask, not imposing their own perspective, changing their own viewswhen warranted, checking out their own and others' assumptions andremaining calm and confident)

When such conditions prevail:

Humans liberate rationality from its bonds through the collective workof civility. (Shulman and Carey, 1984, p. 518)

Empirical evidence in support of the value of collaborative problem-solvingcan be found in the now extensive body of research on peer interaction (see


Developing Expert Leadership for Future Schools

Webb, 1989, for a review of this research). This research unpacks, in more detail,forms of group interaction most helpful in reaching productive outcomes.Schoenfeld (1989) provides a useful, personal case study of how a collaborativeresearch setting fostered achievements not possiblt for him to accomplish work-ing alone.

Contributing to the Long-Term Growth of- Staff

Another important set of abilities for future leaders to possess, evident in thegroup problem-solving of experts, is contributing to the long-term growth ofstaff. Vygotsky's concept of a 'zone of proximal development' has been used inresearch on peer interaction, to help explain how such interaction may stimulateindividual development in a collaborative setting (e.g., Damon and Phelps,1989). This concept also seems valuable in helping to understand why and underwhat conditions group problem-solving by school-leaders and teachers maycontribute to their long-term growth. According to Vygotsky, an individual'sindependent problem-solving is a function of processes in which they haveparticipated in the past - for the most part, processes involving interaction (orcollaboration) with In this sense, an individual's independent problem-solving capacity, at a given point in time, is an internalization of previouslyexperienced, collaborative problem-solving processes; it is their actual develop-mental level. The zone of proximal development:

is the distance between the actual development level as determined byindependent problem solving and the level of potential development asdetermined through problem solving ... in collaboration with morecapable peers. (Vygotsky, 1978, p. 86)

In the context of school-leader/staff collaborative problem-solving, the long-termgrowth of staff seems likely when, for example:


The process used by the group is actually superior to an individual'sindependent problem-solving and the individual participants recognizethat superiority. (Experts in our study seemed to be achieving this, byplanning carefully in advance for how group problem-solving wouldoccur, by actively facilitating the group's problem-solving, and by anti-cipating constraints)There are opportunities for the group to reflect consciously on theprocess in which they are involved, to evaluate it, and to participate in itsrefinement. (Experts provided for this by explicitly outlining to teachersthe process they planne,l. by planning for follow-up, and by gentlykeeping the group on task)Individual members of the group compare their own independentproblem-solving with the group's processes and identify ways of increas-ing the robustness of their own independent processes. (This was acondition experts in our research appeared not to address, but one that isimportant, nevertheless)


Expert School Leadership in the Swamp

Joseph Schwab's (1983) conception of curriculum deliberation is based on aset of premises very similar to those captured in this discussion of collegialrationality. He also argues that teacher involvement, in the kinds of deliberationswhich he advocates, leads to a form of long-term growth, critical to his concep-tion of teaching as a complex and demanding art. Teachers, from this view, arerequired to decide hundreds of times a day and, as Shulman (1984) notes, theiroptions arise differently every day with every group of students. No theoreticalprinciples or abstract guidelines are sufficient to the task faced by teachers underthese circumstances. What is required, rather, is a deep understanding of theirpurposes and how such purposes may be accomplished flexibly and often oppor-tunistically (what we have been referring to as 'vision'). Such understandingarises through thoughtful and extensive deliberations about the nature of instruc-tion, the school curriculum, and other problem areas that teachers are expected toaddress at higher stages in their development (Fullan and Connelly, 1987, sec alsoChapter 8).

Developing Staff Commitment to, and through, Shared Goals

Future schools seem likely to provide opportunities for teachers to play a moreprofessional role. Trends, for example, toward school-based management andteacher empowerment, as discussed in Chapter 3, can be viewed as having thatresult. As many argue, however, (e.g., Little, 1982, Rosenholtz. 1989, Fullan,199(1) for this role change to result in F etter experiences for students, thisprofessionalism will have to be pursued in a collegial way. School-leaders willneed to be able to develop among staff a strong commitment to shared goals, anda widely shared, defensible vision.

There is much evidence in support of the claim that at least some forms ofinvolvement or collaboration in problem-solving lead to greater commitment byparticipants to implement solutions arising from such problem-solving (e.g.,Ealing and Jago, 1988). Under what conditions of involvement does commit-ment arise and why? Conventional wisdr.,n has it that it is simply the participa-tion in the solution process that leads to greater commitment to the solutionitself. Though this is no doubt true, we sec that the reason for the increasedcommitment is due to the concept of shared goals. Increased commitment as aresult of shared goals occurs because: (a) an individual's goals arc a primemotivator of their behavior; (b) an individual's goals are arrived at, in part,through social interaction; (c) certain characteristics of goals are more motivatingthan others; and (d) some forms of social interaction produce goal characteristicsbetter than others, in particular, forms of social interaction which lead to sharedgoals.

When individuals commit themselves to explicit goals, perceived nega-tive discrepancies between what they do and what they seek to achievecreate dissatisfactions that serve as motivational inducements for change... Once individuals have made self-satisfaction contingent upon goalattainmcnt, they tend to persist in their efforts until their performancesmatch what they arc seeking to achieve. (Bandura, 1977, p. 161)



Developing Expert Leadership for Future Schools

In our research on expert problem-solving in groups, school-leaders helped makegoals more explicit by sharing their own goals for problem-solving and interpret-ing problems in relation to the larger mission of the school.

Certain characteristics of goals contribute to their role in motivating be-havior. Relatively explicit goals provide a clearer basis for self-evaluation than doambiguous goals. Moderately difficult goals are more motivating than thosewhich seem trivial because of their simplicity, or those which seem unrealisticbecause of their excessive difficulty. Experts, in our research, seemed to assistwith this by encouraging staff discussion of goals and, as already noted, inter-preting problems in relation to the larger vision and mission of the school. Inaddition, relatively immediate or proximal goals (or subgoals) serve as greaterstimulants to action than do remote goals, especially when there are competingdemands on one's attention.

Goals are actively constructed by the individual through social interaction.The nature of such interaction is substantially influenced by the context (orculture) in which people find themselves. Rosenholtz's (1989) research providesevidence of this influence on teachers. Her study showed that in school contextswhich she described as 'static' (dominated by norms of self-reliance and isola-tion), teacher's goals were, for example, idiosyncratic, focusing on maintenanceactivities and interaction with others about social issues or discipline problems.School contexts described as 'moving', in contrast, appeared to foster sharedgoals focused on student learning and interaction with others about instructionalimprovement. In 'moving contexts':

principals interacted with teachers to shape their school reality, to con-struct school traditions ... goals about the importance of students' basicskills came to be commonly shared. (Rosenholtz, 1989, p. 39)

Shared goals, in Rosenholtz's formulation, constitute the initial foundation onwhich to build teacher conunitment, certainly about instructional practices, andfor the confidence to participate in furthcr collaboration. Experts in our researchmanifested a strong concern for arriving at goal consensus.

Collaboration with one's colleagues seems likely to generate not only sharedgoals but also goals which have highly motivating properties. Interaction requiresone to put one's purposes in words and to be clear enough to explain one'spurposes to others. Furthermore, the public nature of such interaction createspressure to set goals which seem worthwhile to others and therefore not likelytrivial. Continuous interaction about shared goals supplements, through theevaluation of others, one's own evaluation of discrepancies between performanceand desired achievement. Finally, because they arc worked out in a deliberativemanner (with thc aid of others), such goals are less likely to be remote orunrealistic.


Fcwer things arc more predictable in the lifc of leaders of future schools than thcconsistent presence of swampy problems. Many of the swampy problems facingcurrent school-leaders will be on the high ground for future school-leaders. So,


1 14

Expert School Leadership in the Swamp

knowing what they are is not of much help. But the processes used by today'sexperts are of significant assistance to future school-leaders. Learning from andabout these processes is a useful way of reducing the time and effort required fordeveloping expertise. It also seems clear from the research reviewed in thischapter that expert group problem-solving processes serve double duty. On theone hand, they solve the immediate swampy problem. And on the other, theycontribute to the long-term capacity of the school to solve future problems. Thisis 'transformational leadership' the empowerment of others which we iden-tified as one of the three components of the leadership problem, ;1, Chapter 1, ofspecial consequence for future school leaders.

I i 97

Chap ter 7

The Special Role of Values inSchool Leadership

About the nature and role of values in solving school-leader problems, there isconsiderably more heat than light. Values, somehow, ought to be important.Social psychology, for exampk, has long offered impressive empirical supportfor their role in peoples' thinking and problem-solving (e.g., Rokeach, 1975).Furthermore, students of educational leadership and administration acknowledgesuch importance in their theoretical reflections. 'Values are central to educationaladministration', notes Wil lower (1987, p. 17) in the sense that administratorsmake choices among competing values and consider the desirability of alternativecourses of action on a daily basis. But as Greenfield has persistently pointed out,the empirical study of administration has traditionally Ignore[d] value and senti-ment as springs of human action' (1986, p. 59).

The research on which this chapter is based was stimulated indirectly bysuch evidence, unrelated to school leadership, as Rokeach's (1975). It was alsoencouraged by theoretical arguments of the sort offered by students of educa-tional administration like Willower and Greenfield. Research describing expertproblem-solving in the swamp, summarized in Chapter 6, created a much moredirect stimulus for this work, however. Whether solving problems alone or ingroups, this research points to substantial differences between expert and non-expert school-leaders in the role played by their values. Such values also appearedto be pervasive and critical: as we noted, for example, in the discussion ofcognitive flexibility in Chapter 6, expert school-leaders 'display extreme cognit-ive flexibility, but they arc always guided by a coherent set of values'. Giventhcse starting points for examining the nature and role of values more carefully,this chapter addresses four sets of questions. The first question is whetherempirical evidence from school-leaders supports implications regarding the signi-ficance of values in their problem-solving and, if so, what role do such valuesplay? A second set of questions is aimed at morc specifically mapping thc terrainof school-leaders' values: Which values are used (or come to the surface) duringthe process of problem-solving and do some appear more frequently than others?

If, as Greenfield (1986) suggests, values arc 'springs of human action',different levels of school-leaders' expertise (e.g., as in Chapter 5) may be ex-plaMed, in part, by adherence to different values. Indeed, such differences may beas important in accounting for variations in school-leaders' problem-solving as isknowledge and skill, the most popular source of explanation. Our third set of

98 11 G

The Special Role of Values in School Leadership

questions is about the relationship between values and different levels ofproblem-solving expertise. Finally, we ask how school-leaders resolve valueconflicts this, in recognition of the competing values typically encountered,especially in solving the kind of swampy, 'people problems' evident in theresearch reported in Chapter 4 and by others (e.g., Strike, Soltis and Haller,1989).

Three of the studies used to answer these questi-ms were designed andconducted in parallel and largely independent of one another. They were guidedby different frameworks and used different methods of data collection. Thefourth study was specifically designed with the results of the first three in mind.Its purpose was to develop a conception of the nature and role of values thatcould account for the results of the previous studies and to further test the claimsresulting from those studies.

The next section of the chapter briefly summarizes the theoretical frame-works and data-collection methods used in each of the four studies: more detailsare provided on these matters in this chapter than is our practice in other chaptersbecause we believe it to be more crucial to an appreciation of the results. This isfollowed by a scction in which the combined results of all studies are broughtto bear on each of the four sets of questions of interest in the chapter. Asubsequent section explores the implications of the studies for the leadership offuture schools.

Frameworks and Methods: Four Studies


Each study conceived of a value as: 'a conception, explicit or implicit, distinctiveof an individual or characteristic of a group, of the desirable which influences theselection from available modes, means and ends of action' (Hodgkinson, 1978,p. 121). Embedded in this definition are attributes of values also evident in thework of Rokeach (P)75), Kluckhon (1951). Smith (1963) and Williams (1968).That is, a value:

is an enduring belief about the desirability of some means; andonce internalized, a value also becomes a standard or criterion for guid-ing one's own actions and thought, for influencing the actions andthought of others, and for morally judging oneself and others

A person's value system, Rokeach (1975) suggests, is a learned system of rules formaking choices and for resolving conflicts. While based on this common defini-tion of a value, the four studies thought about value types and relationshipsdifferently. Studies One and Two were guided by previously developed valuesframeworks: those of Christopher Hodgkinson (1978) and Clive Beck (1984a)respectively. Study Three 'discovered' values among a set of data, thought to beprimarily concerned with purely rational aspccts of principals' problem-solving.These data were largely useful in clarifying the role that values play in suchproblem-solving. The final study was guided by an original values' frameworkconsistent with results of the previous three studies.


Developing Expert Leadership for Future Schools

Methods used to coll.xt data about school-leaders' values involved, in oneway or another, having school-leaders t.. - out loud about how they would solvea problem. Sometimes this was a problem thcy had solved in the past, sometimesa simulated problem given to them. Sometimes they were asked to listen to atape recording of themselves and their staffs, engaged in group problem-solvingand tell the researcher what they were thinking about at the time. In combina-tion, these different methods serve to control for most of the sources of invalidityusually attributed to verbal reports (Ericcson and Simon, 1984, Nisbett andWilson, 1977) including the distorted reporting of cognitive processes, incom-pleteness of description, and failure of respondents to rely on memory or to relyonly on what can be retrieved from long-term memory.

Study One: Begley (1988)

School-leaders' decisions to adopt and promote the use of computers in theirschools provided the context for this study. Its purpose was to learn more aboutthe nature of values related to such decisions, the relative influence of values incomparison with other factors on these decisions, and the relationship betweenschool-leaders' orientation to their role and the values which they used indccision-making.

Hodgkinson's (1978) conception of value types and relationships served as aframework for collecting data about schooi-leaders' values (Evers, 1985, offers acritical analysis of the framework). Three categories of values are included inHodgkinson's framework:

1. Transrational values grounded in principle.2a. Rational values based on an individual's assessment of consequences,

the attainment of what is perceived as right./b. Rational values based on an individual's assessment of consensus,

again, the attainment of which is perceived as right.3. Subrational values related to personal preferences or what is per-

ceived as good.

Type 3 values represent an individual's conception of what is 'good'. Suchvalues are grounded in affect or emotion and constitute the individual's prefer-ence structure. They are self-justifying and primitive. Each of the two remainingcategories of values describes a 'rightness' that, according to Hodgkinson, ishigher than the one below it. Type 3 values, unlike the others, represent what is'good' as opposed to 'right'.

Type 2 values are subclassified: Type 2b values attribute 'rightness' toconsensus or the wiil of the majority in a given collectivity. Type 2a values define'rightness' in relation to a desirable future state of affairs or analysis of the conse-quences entailed by the value judgment. Type 2 values, as a whole, are rational;Type 3 values subrational; and Type 1 values are transrational.

Hodgkinson argues that Type 1 values are superior, more authentic, betterjustified or more defensible than the other two types. Indeed, use of these 'sacred'values in decision-making, according to Hodgkinson, is the hallmark of theethical educational leader. Such a leader: 'seeks to increase his own degrees of



The Special Role of Values in School Leadership

freedom (a Type 1 value) and the degrees of freedom of those who functionunder his aegis' (1986, P. 8). However, Hodgkinson (1978) also claims that values

tend to lose their level of grounding with time, thereby reducing their authentic-ity or their force of moral insight. He is critical, for example, of what he sees asthe widespread use of Type 2 rational values in administration and attributes it to

a positivistic, impersonal view of organizations and a natural desire to avoid the

messiness and unpredictability associated with use of other types of values. Thistendency toward rational values is greatly reinforced by the characteristics ofcontemporary culture, according to Hodgkinson.

Interviews conducted with fifteen elementary school principals (abouttwo-thirds of the principals in one central Ontario school system) provideddata for Study One. Each interview lasted from one and a half to three hoursand was tape recorded and transcribed.

Study Two: Campbell-Evans (1988)

In addition to the purposes for Study One, Campbell inquired about howprincipals responded to problematic situations in which courses of action prop-osed by others (e.g., senior administrators) conflicted with their values. Thevalues' framework used in this study was developed by Beck (1984a,b,c). It isbased on the premise that a fairly common set of universal values exists. Priori-ties and emphases may shift over time and with respect to specific circumstances.Nevertheless, a set of 'Basic Human Values' can be identified, since values arise

from need and many individuals have similar needs. These values arc `part ofhuman nature and the human condition' (Beck, 19846, p. 3) and include forexample: survival, health, happiness, friendship, helping others (to an extent),

respect for others, knowledge, fulfillment, freedom, and a sense of meaning in

life. Some of these values are means to others but this cluster of basic humanvalues, according to Beck, is mainly ends-oriented. Furthermore, these values are

interconnected and are continuously being balanced (or traded off) with others.A sense of fluidity, openness and flexibility exists within this formulation.

In addition to basic human values, Beck (1984a, p. 3) identifies four othercategories of values: moral values (e.g., carefulness, courage, responsibility);social and political values (e.g., tolerance, participation, loyalty); intermediate-range values (e.g., shelter, entertainment, fitness); and specific values (such as a

car, a telephone and a high school diploma). According to Beck's conception

there are no absolute values. He emphasizes the importance of regarding valueswithin their own system rather than in isolation. Values are both means andends. Viewing a value as merely a means is to deny its intrinsic worth. Viewingit merely as an end is to make it into an absolute. Even the `Basic Human Values'

category forms a set `each of which has considerable importance in itself but must

also be weighed against other values' (Beck, 1984c, p. 4).All eight elementary and junior high school principals in a small urban

school district in Alberta provided data for Study Two. These data were collectedin three phases. Phase One included retrospective, audio-taped interviews withprincipals, concerning two or three prior but recent decisions considered 'impor-

tant' by the principals; these data were analyzed before the second phase of datacollection was initiated.


Developing Expert Leadership for Future Schools

Phase Two required principals to think aloud as they responded to fivesimulated decision problems presented by the researcher; these responses werealso audintaped and transcribed. At the completion of think-aloud responses,principals were asked to react to the 'accuracy' of an analysis of their responses inPhase One. Each principal was given a written report of the identification andprioritization of values revealed in their Phase One interviews and their reactionwas requested.

Phase Three consisted of an interview designed to identify the level ofexpertise on the high ground of each principal.

Study Three: Leithwood and Stager (1989)

Study Three, other parts of which were described in Chapter 6, did not beginwith an interest in school-leaders' values. Rather, as described in Chapter 6, itsexplicit objectives were to identify the components of school-leaders' problem-solving processes and to explore similarities and differences between expert andnon-experts in the swamp with respect to such components. As we reportedearlier, however, values emeiged from the results as an important component ofproblem-solving and one within which expert and non-expert school-leadersdiffered markedly. This study was most helpful in clarifying the role of values intheir problem-solving.

Study Four: Leithwood and Steinbach (1991) (reported here for the first time)

This stady was intended to test the extent to which some of the results of theprevious three studies could be generalized. It inquired about the nature and roleof values with an additional set of data. The six components of problem-solvingidentified in Study Three provided the overall framework for the study. But toguide questions specifically about the nature of values, Beck's (1984) framework(Study Two) was modified on the basis of what had been learned in the previousstudies, specifically about administrators' values (see Table 7.1). The modifiedframework also incorporated those categories of values in Hodgkinson's (1978)framework shown to be relevant to administrators in Study One.

The first category 'Basic Human Values' incorporates values at the apex ofHodgkinson's hierarchy which he calls principles. These are primarily terminalvalues: they refer to 'end states of existence' (Rokeach, 1975, p. 160). Theremaining categories of values worth striving for are more instrumental innature. They represent preferable modes of conduct although, as Bcck (1984a)warns, the distinction between means and ends is difficult to maintain. Peoples'values act as interdependent systems to influence their problem-solving. Categor-ies entitled `Gereral Moral Values' and 'Professional Values' include norms ofconduct or guidelines for judging the ethics of an individual's actions. 'Profes-sional Values', an addition to Beck's framework, includes values uniquely re-levant to guiding decisions in one's work life; Hodgkinson's (1978) values ofconsequence are included here. As Bayles (1981) suggests, in order for 'Profes-sional Values' to be guides to ethical conduct they must be consistent with andsubordinate to 'Basic Human Values'.


The Special Role of Values in School Leadership

Table 7.1: A classification of values and illustrative statements

Categories of Values Illustrative Statement

Set 1. Basic Human ValuesFreedom



Respect for others


Set 2 General Moral ValuesCarefulness

Fairness (or justice)

Set 3

Consequences ..

Consequences (other)

Set 4 Social and Political ValuesParticipation


Professional ValuesGeneral Responsibility asEducatorSpecific. Role Responsibility


Loyalty, Solidarityand CommitmentHelping others

Staff is not forced to supervise dances bythe Education Act ... I would not force peopleto do thisMost people felt pretty good about thosegoalsI would collect as much information about theprobable suspects as possibleIn a blanket approach you could offend manyfine teachersI don't think you can let an issue like thisdominate a lot of time

ICheckl to indeed see if whether or not wehave a problemMake sure that some people who are a littleunsure of themselves also have anopportunity to speakTheir responsibility is to speak out whenvandalism occurs

Your value system is interfering with themandate that we have in educationStaff have to feel they are supported by theofficeKids deserve a certain number of socialeventsThere's an impression that studentsaren't under control

Involve groL, s such as Heads' Council,Special Education, Student ServicesAllow people to get things off theirchests talk about the problems theyperceiveWe [admin. tean-iI have to be seen as beingphilosophically in tuneLet's help each other [school and parent)deal with that child

'Social and Political Values', incorporating Hodgkinson's (1978) values ofconsensus, recognize the essentially social nature of human action and the needfor individuals to define themselves in relation to others to make their livesmeaningful. There is also a close link between the specific values in this categoryand the 'Basic Human Value' of respect for others.'

The categories of values included in Table 7.1 do not include Beck's short orintermediate range values. Such values did not emerge in Studies One, Two orThree as of much ielevance for principals.



Developing Expert Leadership for Future Schools

Data for Study Four resulted from a replication of methods used in StudyThree (ththk-aloud responses to a set of simulated problems), with a group ofeleven expert secondary principals; in this case, they were selected as experts inthe manner also described in Study Three (see Chapter 6).

The next four sections of the paper use these studies as a whole to answer thequestions of interest in this chapter. In each section we also point out thesimilarities and differences between results of these studies and relevant studiescarried out with managers of non-school organizatibns.

The Role and Significance of Values in School-Leaders'Problem-Solving

Three studies helped clarify the role that values play in school-leaders' problem-solving. Study One inquired about the role of values in their decisions about theadoption and implementation of computer technology in their schools. Resultsdemonstrated a pre-eminent role for values in the adoption decision but a muchless important role in the subsequent process of solving the implementationproblem. For that problem, factors identified in school-improvement research(e.g., fit, building user commitment) appeared to be more influential. Theseresults appear to reflect one of two relationships between values and actionspropo. :d by Hambrick and Brandon (1988): a direct relationship that is,dominant or very strong managerial values are thought to be capable of dictatingbehavior without any (or much) regard for facts. Indeed, Study One providesespecially relevant support for this role of values since a number of school-leadersin the study decided to adopt computers without knowing what the consequ-ences would be for students or othcrs. This role for values has been termedbehavior channeling' (England, 1967).

Study Two examined the role of both internal influences (i.e., beliefs andvalues) and external influences (e.g., time, money) on school-leaders' problem-solving. Results argue for a more pervasive role for values than did Study Oneby suggesting that values give meaning to potential external influences and act asfilters in determining whether potential external influences will be allowed to beactual influences. So, for example, factual information in a report available toschool-leaders, relevant to some aspect of their work, is more likely to influencethat work if they strongly value 'knowledge' (a 'Basic Human Value') than ifthey do not. These results illustrate the second type of relationship betweenvalues and action suggested by Hambrick and Brandon (1988): this is an indirectrelationship, termed by England (1967) 'perceptual screening'. Values influenceschool-leaders' perceptions of events causing them to attend closely to some andIgnore others altogether. These highly subjective perceptions then lead to action

Finally, Study Three which compared the problem-solving of expert andnon-expert school-leaders, especially in response to ill-structured problems,found that, for experts, explicit values acted as substitutes for knowledge. Anill-structured problein, by definition, is one about which the solver possesseslittle problem-relevant knowledge. Unlike non-expert school-leaders, expertswere relatively clear about their principles and values, and so were able to makeuse of them as guidelines for problem-solving (e.g., 'I may not know exactlyhow to solve this problem but whatever we do, we are going to be open and



The Special Role of Values in School Leadership

honest with everyonc'). This role played by values when there is little problem-relevant knowledge appears to be the same role proposed by Barnard (citedin Hambrick and Brandon, 1988) when managers face the opposite: excessiveknowledge about a problem. Values, he suggests, provide a moral code forsorting a bewildering /oad of information and options that may confront themanager. In the absence of such a code, the manager bogs down.

In sum, then, dominant values appear to play an especially explicit andimportant role (behavior channeling) at key points in the problem-solving pro-cess. But throughout, they also act, more subtly, as perceptual screens fordetermining what aspects of the wider environment will be considered. And theyare substitutes for knowledge (or moral rules of thumb) in the face of novelproblems.

These studies provide considerable support for the theoretical claims thathave been made by Hodgkinson (1978, 1986), Greenfield (1986) and others aboutthe importance of values in school-leaders' problem-solving.

Types of Values and Their Relative Importance

Only Study Three did not inquire about the nature of values used by school-leaders in their problem-solving and the relative importance of such values.Results from Study Four represent the findings of Studies One and Two reason-ably well and are reported first. These results speak to not only types of valuesused by school-leaders and their relative weight but they also explore possibledifferences that might arise, depending on whether a problem is viewed by theschool-leader as a high ground or swampy problem.

Results from Study Four indicated that 'Professional Values' was the mostfrequently cited category of values; the least frequently cited was 'General MoralValues', and 'Basic Human Values' and 'Social-Political Values ranked secondand third.

Within categories, there was consistency in the specific values cited whethermost or least clear problems were being addressed. 'specific role responsibility','consequences for one's immediate clients' (e.g., students) and 'knowledge'appear to be the dominant specific values cited by school-leaders in theirproblem-solving. Also cited frequently in relation to most and least clear prob-lems were 'general responsibilities as an educator', 'participation' and 'respect for

others'.Consistent with findings in Study Three, values emerged more frequently in

principals' responses to least clear (total frequency = seventy-four) as comparedwith most clear (total frequency = forty-nine) problems. This contrast is, in fact,

greater than it appeared in our data because results were unavailable for oneprincipal's response to the least clear problems. Adjusting for these missing datasuggests that values were mentioned about 65 percent more frequently in re-sponse to least clear as compared with most clear problems.

Most of the differences between the results of Study Four and the results ofStudies One and Two were due to differences in the frameworks used for coding.Study One, guided by Hodgkinson's framework, found values of consequenceand consensus to be used most frequently by principals. Study Two attributed


Developing Expert Leadership for Future Schools

greatest weight to Beck's 'Social and Political' category of values and found thespecific value 'Responsibility' to be used quite frequently. 'Responsibility' and'consequences' values are part of the most frequently used category of 'Profes-sional Values' in Study Four; 'consensus' is part of the category 'Social andPolitical' values common to Study Two and Four frameworks ranked first inStudy Two and third in Study Four (although only marginally less frequentlymentioned than the second ranked category). 'Respect for others' and 'know-ledge' appear to receive comparable frequencies of mention in Studies Two andFour; they were not a part of the Study One framework.

In sum, results from the three studies concerned with types of values andtheir relevant importance were highly consistent. This consistency appeared inspite of the use of somewhat different data-collection procedures, school-leadersdrawn from different geographical regions and school levels (elementary andsecondary), and school systems varying widely in size and expectations for theirprincipals. Even though the size of the combined sample of principals studiedwas small (thirty-four), such consistency adds to the confidence one can have inthe robustness of thc results.

Based on their review, Hambrick and Brandon (1988) concluded that six setsof values encompass the results of efforts, to date, to identify managerial values.At least some aspects of four of these six sets appear in our values framework (seeTable 7.1):

Collectivism: includes at least 'Respect for Others' (in set 1) and prob-ably 'Fairness' (in set 2)Duty: includes 'General Responsibility as Educator' and 'Specific, RoleResponsibility' (in set 3) 'Loyalty and Solidarity' and 'Sharing' (in set 4)Rationality: includes 'Knowledge' (in set 1)Materialism: may be aspects of 'Happiness' and 'Freedom' (in set 1) forsome people

Our classification scheme does not directly reflect Hambrick and 13randon's(1988) 'novelty' (to value change, the new, the different) category althoughHodgkinson's personal preferences seem to be this. Nor is their evidence amongschool-leaders in our four studies of the Hambrick and 13randon's category'power' (to value control of situations and people). Parenthetically, Leithwoodand Steinbach (1991) did find evidence of a power/control value among chiefeducation officers whose roles more closely approximated the executive rolesincluded in much of the research reviewed by Hambrick and Brandon. Perhapsposition in thc hierarchy has some relationship to leaders' values.

Comparing the Hambrick and Brandon (1988) results with those fromStudies One, Two and Four suggests that a large core of values are shared notonly among school-leaders but across managers in many different types oforganizations. This comparison also suggests some differences, with school-leaders showing greater evidence of such social values as 'participation' and lessevidence of valuing 'novelty' (or personal preferences). This combination ofdifferences supports a relatively conservative orientation to change by presentschool-leaders: conservative, because opportunities for radical change arc bluntedby others through participation, and the value 'novelty' is more weakly held.


The Special Role of Values in School Leadership

Relationships Between School-Leaders' Values and Expertise

Studies Two and Three examined the relationship between school-leaders' valuesand their levels of problem-solving expertise (see Chapters 5 and 6).

School-leaders included in Study Two (using Beck's framework of values)divided themselves evenly between two levels of expertise on the high ground.Half the principals appeared to be humanitarians (Level 2) and half were programmanagers (Level 3) (Fig. 5.1). Evidence from this study suggested that, whileboth sets of school-leaders shared a common core of valus.., a strong relationshipexisted between levels of high ground expertise and some specific values, mostnotably within the category 'Basic Human Values', mentioned most frequently.'Respect for others' was the most frequently mentioned 'Basic Human Value' byschool-leaders labelled humanitarians. For the more expert program managers'knowledge' was the most frequently mentioned 'Basic Human Value'. Withinthe category of 'Social and Political Values', both sets of school-leaders frequent-ly mentioned 'participation', but humanitarians made greater mention also of'sharing'. 'Responsibility' was the moral value mentioned most by both sets ofschool-leaders. Humanitarians also identified 'carefulness' as a value in this

category.School-leaders in Study Three were divided into an expert group and a

non-expert group. For the most part, the expert group engaged in Level 4 highground problem-solving (see Fig. 5.1). Most non-experts were humanitarians.Comparing the responses of these two groups to high ground problems revealedfew differences with respect to principles and values. Three differences wereevident, however, in responses to swampy problems. As compared with thenon-expert group, experts more frequently drew upon principles and values intheir problem-solving. Given current expectations for the role, the principles onwhich the experts drew also seemed more defensible: for example, greater atten-tion was given to consequences for students. Finally, the expert group morefrequently relied on 'Specific Role Responsibility' as a value in approachingswampy problems: this finding is consistent with evidence collected from chiefeducation officers (Leithwood and Steinbach, 1991) and expert secondary princip-als (Leithwood and Steinbach, in press b).

Our research suggests, in sum, that school-leaders rely on a common core ofvalues in their problem-solving, independent of their levels of expertise. This isthe case, in particular, with 'General Moral Values' and 'Social and PoliticalValues'. Since education is indeed a 'moral' enterprise and therefore likely toattract people with similar 'Basic Moral Values', this is not surprising. Further-more, as a 'boundary spanner', principals are regularly in communication withmany different groups of people and spend as much as three-quarters of theirtime in personal communication (Martin and Willower, 1981). It would bedifficult to avoid seriously acknowledging the influence of 'Social and PoliticalValues' in such an environment.

But there does appear to be evidence of relationships between patterns ofpractice and both 'Basic Human' and 'Professional' values. For example, thehumanitarians' preoccupations with school climate and interpersonal relationshipsseems consistcnt with the frequency of mention of the value of 'sharing'. Similar-ly, program managers' concerns for the quality of the classroom learning


Developing Expert Leadership for Future Schools

environment seems consistent with the stress they give to 'knowledge'. Theserelationships provide tentative support for school-leaders' values as a partialexplanation for or a variable interacting with their levels of expertise.

More specifically, results concerning the dominant values of effective school-leaders are intriguing in light of Hambrick and Brandon (1988). Their reviewstimulated the suggestion that the more managers value rationality (what wehave called 'knowledge') the more their other values will operate through percep-tual screening: this also appears to be the case as managers increasingly value duty(our 'responsibility' values). Since our evidence depicts the most expert school-leaders adhering strongly to both sets of values, a substantial direct influence ofvalues on action seems likely. This assertion assumes considerable discretion forschool-leaders to act in concert with their values since organizational constraintson school administrators' actions will blunt the influence of values on action.Indeed, as school-based management creates more discretion for school-leaders,their values are likely to become an increasingly productive focus of attention,especially for purposes of school-leader selection.

Resolving Value Conflicts

This issue was addressed by Studies Two and Four. As part of Study Four,school-leaders were asked to: (a) describe a problem in which they had beeninvolved which had a great deal of value conflict, (b) indicate what were thecompeting values, and (c) outline how they had dealt with the conflict.

Our understanding of responses to this question indicates that school-leadersencountered two types of value-related conflicts. One type involved competitionbetween two or more values for recognition in the formulation of a solution.Such competition took three different forms:

Value conflicts between two or more people other than the principalwith the principal acting as mediator (e.g., enrichment teachers andregular classroom teachers disagreeing about the meaning of treatingstudents fairly in the context of an 'honor week');Value conflicts between the principal and other staff members (e.g.,general moral value of principal with respect to adultery in conflict withthe values of two married staff members having an affair with oneanother; the principal was also quite concerned about the consequencesfor students);Value conflicts concerning thc principal alone (e.g., principal 'caught'between the need to act quickly to remove from the classroom a teacheraccused of inappropriate behavior with an older student, and the value offair treatment and due process for the value of fair treatment and dueprocess for the teacher)

A second source of conflict for school-leaders was between a set of valuesstrongly held by them, and their actions. This conflict was usually experienced asan inability to act in a manner consistent with values held. One school-leader, forexample, held a strongly nurturant attitude toward his teachers and experiencedconsiderable conflict in bcing unable to effectively counsel one teacher toward aconsistent and productive career plan:



The Special Role of Values in School Leadership

It's a value conflict for me because I think she's making all the wrongdecisions and it breaks my heart and I cannot do anything about it.

Principals also used two distinct processes for attempting to resolve valueconflicts. The first type we called 'deep and strong' because it appeared to beanalogous to the solution processes of expert school-leaders in response to othersorts of swampy problems. This process, used by about half of the school-leadersto resolve conflicts between competing values, included: taking considerable careand effort in the early stages to clarify the nature of the conflict; satisfyingthemselves that the problem could not be usefully interpreted as involvinganything but serious value conflicts (i.e., avoiding such an interpretation wherepossible); and clarifying for themselves their own priority ariong the competingvalues. As part of the process, school-leaders relied on fo mal, organizationalprocedures, where appropriate, for esolving value confl.cts (e.g., teacher-dismissal procedures) once the conflicts were clarified. Other less formal butsystematic procedures such as information collection, collat. oration with others,and consensus-reaching techniques were also used. win... procedures failed,as they sometimes did, school-leaders capitalized on unanticipated opportunities(e.g., a parental complaint about a teacher).

The second type of conflict-resolution process used by school-leaders, welabelled 'surface and weak'. Only one school-leader used a version of this processto resolve a conflict between two values. Three school-leaders, however, used itin an effort to find a course of action consistent with their values. In no case didthis process result in a solution satisfactory to the school-leader. The process wasone in which school-leaders often sought out others' interpretations of thc con-flict and consulted with others about solutions. Nevertheless, the cause of theconflict usually remained unclear (e.g., a teachers' erratic behavior, a teacher'slack of interpersonal skill) and possible courses of action were considered andtried sequentially (and, in one case, only half-heartedly). School-leaders using thisprocess seemed less clear about their own relevant values and had fewer existingprocedures to call on as supports for their own actions.

Study Two approached principals' resolution of value conflicts from a differ-ent perspective than Study Four. As a result, the two sets of research results arebest viewed as combining to reveal a larger proportion of a still incompletepicture, rather than two separate snapshots of the same completc picture. Where-as Study Four included school-leadcrs' responses to value conflicts arising fromproblems within the school, Study Two examined conflicts created through theimposition of policy from outside the school. School-leaders were presented witha simulated demand from their districts to implement a policy which conflictedwith some of their central values. This type of value conflict constrains alterna-tives for acting more severely than does the value-conflict situations of interest inStudy Four. It required school-leaders to weigh one specific value 'respect forauthority' (Hambrick and Brandon's 'duty') against a range of personally heldvalues which varied across school-leaders in the study.

Seven out of eight school-leaders in the study clearly indicated that it wasthcir duty to implement the policy as written. Comments such as 'you stillfollow the policy because it's policy', 'you are bounded by Boald policy', and'as an administrator, I must administer the policy', reflect the degree to whichthey felt committed to the implementation of official policy. Their individual

127 109

Developing Expert Leadership for Future Schools

preferences took second place to their 'respect for authority' principle whichcompelled them to follow the policy as an initial action. In this study, 'respect forauthority' is difficult to separate from 'specific role responsibility' in Study Four.One school-leader recalled a conflict which was the result of staff having no rolein policy development. For another, a situation was perceived as unfair becauseof the demands placed on staff. Conflict, in a third example, arose from what theschool-leader perceived as insufficient time available to meet the requirements ofnew policy.

hi spite of whatever differences school-leaders had with the policy, however,they initiated its implementation. Then they began to deal with the conflict. Acommon, first step of implementation initiated a series of events. While actingwithin the policy, all eight principals indicated that they would 'do something',that is, some action would be taken. Action was first an expression of concern tosignificant others about the policy's content which subsequently involved theseindividuals or groups in the conflict situation. In most cases this expression ofconcern involved the immediate school staff initially and , then expanded toinclude the broader community of other school-leaders and/or the superinten-dent, or the elected trustees. The opinion of other school-leaders was sought infive of eight cases through the forum of the school-leaders' meeting. A desire toincrease the understanding of staff was expressed by three principals. This ex-pression of concern took shape through formation of school committees, work-ing from within the policy toward change, working with the policy, workingwith staff, preparing a presentation for the Board, and letters to the superinten-dent. In sum, they dealt with the conflict by working with colleagues to changeit. These actions reflect other important values which came into play for eachschool-leader. Their value structures readjusted and the values of participationand sharing, for example, became apparent.

As a whole, these studies suggest that when school-leaders encounter valueconflicts their responses are more productive when the conflict itself is treated as aproblem and subjected to the same (deep and strong) processes that would beused with other types of swampy problems. It also appears that even whenconflicts arise between two unequal values (e.g., where one of the two valuesclearly carries more weight in the school-leader's value system) school-leaders donot simply choose one and reject the other. Rather, they search for compromises.

Such processes for resolving value conflicts are similar to conclusions drawnby Toffler (1986) in her study of twenty-one non-school-based managers. Herresults attribute importance to managers defining the elements of the conflict,assessing their own 'specific role responsibility' and using their imagination toidentify a key factor on which the dilemma hinges or developing a mechanism toturn that key factor. When a key factor cannot be found, managers give greaterweight to one value but also respond to the other as fully as possible.

Summary and Implications for Developing Leaders forFuture Schools

Four studies reported in this chapter have been used to answer questions aboutthe special role that values play in solving swampy school-leadership problems.



The Special Role of Values in School Leadership

Concerning their role and significance, first, these studies depicted values aspervasive in the problem-solving of school-leaders through their direct stimula-tion of action and their roles as perceptual screens and moral codes or substitutesfor knowledge in response to ill-structured problems. Second, with respect to thetype of values held by school-leaders and their relative weight, the studiessupported the comprehensiveness of a four-fold classification of values including'Basic Human', 'General Moral', 'Professional' and 'Social and Political' Values.Within these categories, school-leaders cited most frequently the specific valuesof 'role responsibility', 'consequences' for students and 'knowledge'.

The four studies also pointed to a plausible set of relationships betweenvalues and school-leaders' levels of expertise, although principals did share acommon core of values. Greater emphasis was placed on social values (e.g.,'sharing') by those engaged in moderately expert (humanitarian) problem-solving. 'Knowledge' and 'role responsibilities' were dominant values forexperts.

Finally, conflict-resolution strategies used by school-leaders in the face ofcompeting values were of two types. The most successful strategies, 'deep andstrong', conceptualized value conflicts as problems in their own right and em-ployed a set of deliberate problem-solving processes to resolve them. Less suc-cessful were 'surface and weak' strategies which gave short shrift to probleminterpretation or clarifying the source and nature of the conflict, and used asequential, trial-and-error procedure for determining the consequences of alternat-ive solutions.

A number of implications for developing leaders of future schools arcinherent in the results of the four studies.

Use l'alues as Criteria during the Selection of School-Leaders

Although school-leaders' values may change, it is not likely that they will changequickly or easily (Hambrick and Brandon, 1988). Beliefs and attitudes will needto change in order to stimulate and reinforce value change. The challenge ofchange in school-leaders' values, to serve better the interests of the school, is bestavoided, where possible, by school-leader selection processes which collect evid-ence about: the nature of applicant values, the degree to which applicants are clearabout their values, and how well they are able to resolve value conflicts. Thispossibility is examined more fully in Chapter 14.

Redesign School-Leader Socialization Promses

Values develop more through the usually lengthy periods of informal socializa-tion than through formal programs. But the socialization of administrators inmost school districts is largely left to chance. More planful efforts to redesignsocialization experience which will reinforce values important to future schoolsare called for. Chapter 10 provides information useful when pursuing this im-plication further.


Developing Expert Leadership for Future Schools

Provide the Discret;on Needed to Act on One's Values

When satisfactory steps have been taken to hire and/or develop school-leaderswhose values are consistent with the needs of future schools, it will be importantto give them the discretion to act on the basis of those values. This may turn outto be an argument for more school-based management in one form or another. Itis, at minimum, a call to design school districts in a way that support andenhance the work of expert school-leaders. What this might entail is exploredfurther in Chapter 15.


These three implications for developing leaders of future schools seem warranted,given the limited research on which this chapter was based. Nevertheless, muchmore research is clearly called for. It already seems clear that several of thequestions raised in this chapter will be relatively easy to answer and others muchmore complex. For example, values clearly play a significant role in problem-solving. But the nature of that role is by no means fully captured by our data. Inorder to build on our data in subsequent research, an explicit theory about therole of values in managerial thinking and action, such as Hambtick and Bran-don's (1988), would be useful: their hypotheses offer productive starting pointsfor subsequent inquiry.

Perhaps the most theoretically interesting and practically useful focus forsubsequent study concerns the resolution of conflicting values. Such conflicts area part of the everyday work world of school-leaders and likely to remain so. Yetthe few current texts intending to offer disciplined advice to leaders (e.g., Strike,Soltis and Haller, 1989, Taller, 1986) fall short of accomplishing that goal.Research describing an array of strategies used by expert school-leaders forresolving value conflicts would be helpful for leaders of future schools.

Pursuit of these implications for practice and research is likely to be achallenging but worthwhile business. As Hambrick and Brandon note:


to study executive values is to delve into the murkiest of organizationalphenomena. Yet the role of values in influencing organizational pro-cesses, membership and outcomes is enormous. (1988, p. 3(I)


Chapter 8

Teacher Development:A Central Problem for Leadersof Future Schools1

Transformational leadership, as it was described in Chapter 1, involves develop-ing the capacity of organizational members: the capacity to meet both the im-mediate and long-term challenges involved in moving toward a vision of thefuture widely shared by members of the school. As Schlecty notes:

the key result is the growth and development of others; for the [trans-formational] leader, the goal is to help others succeed. (1990, p. 105)

The most promising focus for the developmental effort of school-leaders is theteaching staff. To the extent that school-leaders solve the problem of contribut-ing to teachers' growth, they can be viewed as exercising transformationalleadership. Having devoted most of the previous four chapters to the processesused by school-leaders to solve problems, in general, we now turn to one of themost important specific problems facing leaders of future schools.

The chapter is intended to accomplish two purposes. First, by providingone brief synthesis of existing knowledge about the nature of teacher develop-ment, we make the problem, at least for the reader, less swampy. This seemsworth doing in the face of the continuing debate about the feasibility of an'instructional leadership' role for the principal (e.g., Rallis and Highsmith, 1986,Gersten et al., 1982). Using Baird's (1983) definition of a problem, as explained inChapter 4, our description of teacher development simplifies problem-solving forthe school-leader by adding clarity to the 'givens' and 'goals' of the problem. It isdifficult to contribute to teacher development in the absence of a clear image ofwhat such development looks like. Such an image ought to be part of theschool-leader's overall vision for the school.

The second purpose for the chapter is in response to those school-leaderswho, while acknowledging their responsibility for teacher development, do notfeel that it is a problem they are capable of addressing adequately. It is, in theirview, excessively swampy because of a perception of their jobs as fast-paced,hectic, unpredictable, interpersonally intense, and sometimes consumed byseemingly trivial but pressing 'administrivia'. In such a context, many school-leaders believe they do not have the opportunity to contribute to teacher develop-ment. The second part of the chapter identifies a small number of guidelines forfostering teacher development accepting this perception of thc job. Indeed, we


Developing Expert Leadership for Future Schools

make the case that the job is ideally suited for assisting in teacher development,although commonly held misconceptions about productive forms of such assist-ance belie this claim. Again, with reference to Baird (1983), our identification ofguidelines for fostering teacher development makes problem-solving for school-leaders less complex. It does this by contributing specificity to steps or actionsschool-leaders can use to overcome constraints to teacher development. Thishelps transform the given or current state (lower levels of teacher development)into the goals or desired state (higher levels of teacher development).

The Nature of Teacher Development

This section of the chapter synthesizes evidence from three distinct areas ofresearch to build a multi-dimensional description of teacher development. It isoffered to school-leaders as an aid in reflecting upon and possibly making morerobust their own views of such development. Table 8.1 describes three dimen-sions of teacher development with which school-leaders have an opportunity toassist: development of professional expertise, psychological development, andcareer-cycle development. Each of these dimensions reflect quite different lines ofinquiry about teacher development.

Development of Professional Expertise

The dimension of teacher development with the most obvious consequences forclassroom, school, and district improvement is identified in Table 8.1 as 'De-velopment of Professional Expertise'. It is through such expertise that teacherscontribute directly to the growth of students (amount learned, range of outcomesachieved and ranges of students who benefit from instruction). Six stages ofdevelopment are included in this dimension. Stages One through Four are con-cerned with teachers' classroom responsibilities; Stages Five and Six explicitlyaddress the out-of-classroom and out-of-school roles of the 'mature' teacher.Each of the stages (beyond the first) includes expertise acquired in previousstages. Furthermore, it seems likely that the seeds of expertise in higher stageswill begin to develop quite early, given appropriate, formative experiences.Hence, this conception of growth does not imply restricting teacher experiencesonly to those that will prepare them for the next stage of development. Somepreparation for Stage Six practices might well begin during a teacher's initialentry into the role. Table 8.2 illustrates in more detail those aspects of profession-al expertise likely to be a part of each of the six stages. While others mightdescribe the aspects of expertise in each of these stages differently, there is at leastgood warrant for the substance of Table 8.2. Stages One to Four are based on animage of effective classroom instruction as requiring a large repertoire of instruc-tional techniques. Such a repertoire is reflected, for example, in Joyce and Weil's(1980) twenty-three models of teaching, organized into four 'families' or categor-ies. Expertise, in these terms, increases as teachers acquire greater skill in applica-tion of a given model and as an increasing number of such models are mastered.Teaching, however, involves more than the unthinking application of such mod-els, although 'automaticity' is an important characteristic of expertise in most



Table 8 I Interrelated dimensions of teacher development



autonomous/interdependentprincipled, Integrated

3conscientious, moral.conditional dependence


conformist, moral,negative, independence


self-protective, pre-moral,unilateral dependence



6participating in broadrange of educationaldecisions at all levels

5Contributing to thegrowth of colleagues'instructional expertise

4acquiring instructionalexpertise

3expanding one'sinstructional flexibility

2becoming cJrnpetent inthe basic skills ofinstruction


developing survival skills


5preparing for retirement,focusing

4reaching a professionalplateau

3new challenges andconcerns


stabilizing, developingmature commitment


launching the career



Developing Expert Leadership Jor Future Schools

Table 8.2: Development of profe.,sional expertise

Professional Expertise Dimension Development

1. Developing Survival Skills Partially developed classroom-managementskillsKnowledge aboot and limited skills in use of

several teaching modelsNo conscious reflection on choice of modelStudent assessment is primarily summativeand carried out, using limited techniques inresponse to external demands le g., reporting

to parents); may be poor link between thefocus of assessment and instiuctional goal

2. Becoming Competent in the Basic Well-developed classroom-management skills

Skills of InstructionWell-developed skill in use of several teachingmodelsHabitual application, through trial and error, ofcertain teaching models for particular parts of

curriculumStudent assessment begins to reflectformative purposes, although techniques arenot well suited to such purposes; focus ofassessment linked to instructional goals

easiest to measure

3 Expanding One's InstructionalFlexibility

4. Acquiring Instructional Expertise



Automatized classroom-management skillsGrowing awareness of need for and exigtenceof other teaching models arid initial efforts toexpand repertoire and experiment withapplication of new modelsChoice of teaching model from expandedrepertoire influenced most by interest inproviding variety to maintain student interestStudent assessment carried out for bothformative and summauve purposes; repertoireof techniques is beginning to match purposes,focus of assessment covers significant rangeof instructional goals

Classroom management integrated withprogram: little attention required to classroommanagement as an independent issueSkill in application of a broad repertoire of

teaching modelsInstructional goals, student learning styles,content to be covered, as well as themaintenance of student interests used ascriteria for choice of teaching modelStudent assessment is carried out for bothformative and summative purposes, using awide array of .chniques; program decisionsare informed by assessment and the focus ofassessment is directly linked to the full array of

instructional goals

Teacher Development A Central Problem for Leaders

Table 8 2 (cont

Professional Expertise Dimension Development

5 Contributing to the Growth Has high levels of expertise in classroomof Colleagues. Instructional instructional performanceExpertise Reflective about own competences and

choices and the fundamental beliefs andvalues on which they are basedAble to assist other teachers in acquiringinstructional expertise through either planning,learned experiences such as mentoring ormore formal experiences such as in-serviceeducation and coaching programs

6 Part,cipating in a Broad Is committed to the goal of schoolArray of Educational ImprovementDecisions at all levels Accepts responsibility for fostering that goalof the Education System through any legitimate opportunity

Able to exercise leadership, both formal andinformal, with groups of adultsinside and outside the schoolHas a broad framework from which tounderstand the relationships amongdecisions at many different leveis inthe education systemIs well informed about policies at manydifferent levels in the education system

,rreas of human endeavor. Along with Joyce and Weil (1980), Darling-Hammondet al. (1983). Bacharach et al. (1987) and Shavelson (1973, 1976) and others pointout that deciding which model or technique to apply in a particular situation iscentral to instructional expertise. As teachers develop, their choice of models isbased on increasingly defensible criteria (e.g., instructional objectives vs. need forvariety) and diagnosis of the instructional needs of students.

While the notion of teacher-as-decision-maker appropriately recognizesthe contingent naturc of the classroom tasks routinely faced by teachers, it isnot sufficiently comprehensive to encompass those unanticipated, non-routine'swampy' problems encountered in the classroom from time to time. Schön(1983) depicts the way in which experienced professionals in many domainsthink about and eventually resolve such problems. This involves a processof 'reflecting-in-action' as well as a process of 'reflecting-on-action' in whichthe unique attributes of the setting are carefully weighed and the professional'srepertoire is adapted in response to such uniqueness. These processes are the sameas described for expert school-leaders.

Stages Five and Six acknowledge the roles of teachers in school improve-ment and educational decisions beyond the classroom and school. While suchroles arc by no means new, they have received much greater attention recentlyNet. coaching (Brandt, 1987, Garmston, 1987) and mentoring (Gray and Gray,1985, Wagner, 1985) strategies, for example, assume those aspects of expertiseidentified in Stage Five, as do many of the recent career-ladder programs whichplace teachcrs in the role of evaluators (e.g., Peterson and Mitchell, 1985). Stage



Developing Expert Leadership for Future Schools

Six conceptualizes the mature teacher as one who plays a leadership role, formalor informal, in a variety of contexts both inside and outside the classroom andschool. Teachers, according to this view, share in the responsibility for mostdecisions that directly or indirectly touch on students' experiences. Such a view isconsistent with recent proposals for reshaping teacher education (e.g., Fullan andConnelly. 1987) and for 'empowering' teachers (e.g.. Maeroff, 1988) in theprocess.

Psythological Development

As outlined in Table 8.1 Psychological Development' is a synthesis of threedistinct and independently substantial strands of psychological stage theory:Loevinger's (1966) seven-stage theory of ego development, Kohlberg's (1970)six-stage theory of moral development and Hunt's (1966) four-stage theory ofconceptual development. These three strands of psychological development areboth conceptually and empirically related (Sullivan. McCullough and Stager,1970). The synthesis provided by Table 8.1 is for heuristic purposes and gives arough approximation of how the three strands might intersect in real time.Generally, ego development occurs as a person strives to master, to integrate andotherwise to make sense of experience. Greater ego maturity is associated with amore complex and better differentiated understanding of oneself in relation toothers. Moral development occurs when the bases on which one's views ofrightness and goodness shift from a basis of personal preference toward a basis ofuniversal ethical principles. Finally, conceptual development occurs as one movestoward greater differentiation and integration of concepts, which means a growthfrom concrete toward more abstract thought processes.

Viewing the three strands of psychological development together, as inTable 8.1, provides descriptions of teachers in various stages of growth. A 'StageOne 'teacher has an overly simplistic view of the world and a tendency to seechoices as black or white. Such a teacher believes strongly in rules and roles,views authority as the highest good and most questions as having one answer.Stage One teachers discourage divergent thinking and reward conformity androte-learning (Oja, 1979). Their classrooms are highly teacher directed.

Stage Two teachers (conformists) are especially susceptible to the expecta-tions of others; their wish is to be like their peers and they may hold stereotyped,distrustful views of those outside their immediate group. The classrooms ofcontbrnust teachers are what we think of as 'conventional'. Rules are quiteexplicit and student behavior is expected to adhere to such rules without muchregard for individual differences or contingencies which might justify exceptionsto the rules.

At the third stage of psychological development (conscientious), teachershave become much more self-aware and have developed an appreciation formultiple possibilities in situations (e.g., multiple explanations for student be-havior). Rules are internalized and applied with an appreciation for the need forexceptions to the rules given the circumstances. Teachers at this stage are future-oriented. and achievement-oriented; their classrooms are the product of rationalplanning and a concern f,A- good interpersonal communication.

At the highest stages of psychological development, teachers arc inner-directed but appreciate the interdependent nature of relationships in a social


Teacher Development: A Central Problem for Leaders

setting such as a classroom. In ddition, according to Oja (1981), these teachershave achieved more of a synthesis in their classrooms between an emphasis onachievement and an interpersonal orientation. They are not only able to view asituation from multiple perspectives but arc also able to synthesize such perspect-ives. Teachers at the highest stage understand the reasons behind rules and soLan be wiser in their application; they maintain a broad perspective and are ableto cope with inner conflicts as well as conflicting needs and duties. The class-rooms of these teachers are controlled in collaboration with students and theemphasis is on meaningful learning, creativity, and flexibility. Being more cognit-ively complex themselves, teachers at this stage encourage more complex func-tioning in their students (Oja, 1979, Hunt, 1966).

Career-Cycle Development

The dimension called 'Career-Cycle Development' in Table 8.1 views teachers'careers from a life-cycle perspective. Five stages of development have beenderived primarily from recent research by Huberman (1988) and Sikes, Measorand Woods (1985). The latter work adopted Levinson et al.'s (1978) conceptualiza-tion of life development as a framework. Huberman (1988) and Sikes. Measorand Woods (1985) carried out their research with secondary school teachers inSwitzerland and Great Britain, respectively. Nevertheless, their results appear tobe sufficiently similar and consistent with other research (e.g., Ball and Goodson,1985) to warrant tentative generalization to other contexts and teaching assign-ments in the modified form described in Table 8.1. Our main interest in teachers'career-cycle development is how it interacts with the development of professionalexpertise. More particularly, we want to know what career experiences, at each

stage, seem likely to foster or detract from the development of professionalexpertise.

Stage One, 'launching the career', encompasses up to the first several yearsof the teacher's classroom responsibilities. Sikes et al. (1985) suggest that mostteachers at this stage experience a 'reality shock in coming to grips with problemsof disciplining and motivating students' (p. 17), as well as some degree of cultureshock, the amount of shock depending on the values and perspectives of staff inthe school in which they find themselves. Nevertheless, Huberman's (1988) datasuggest that experiences at this stage are perceived by some teachers as 'easy' andby others as 'painful'. Conditions giving rise to perceptions of easy beginningsinclude: positive relationships with students, 'manageable' students, and a senseof instructional mastery and initial enthusiasm. Painful beginnings are associate.]with: role overload, anxiety, difficult pupils, heavy time investment, close moni-toring, and feelings of isolation in the school. For those who experience suchpain, there may be a protracted period of trial and error in an effort to cope with

such problems.'Stabilizing', the second career-cycle stage, often coincides with receiving a

permanent contract and making a deliberate commitment to the profession Thisstage is characterized by: feeling at ease in the classroom, mastery of a basicrepertoire of instructional techniques, and being able to select appropriatemethods and materials in light of student abilities and interests. Furthermore, atthis stage teachers act more independently, are less intimidated by supervisors,and feel reasonably well integrated into a group of peers. Some teachers at this



Developing Expert Leadership for Future Schools

stage begin to seek greater rf!sponsibility through promotion and/or participationin change efforts.

The stage following stabilization may take several forms. In the main,teachers at this stage tend to be between the ages of 30 and 40 years. As Sikeset al. (1985) point out, their experience is substantial by this point as is theirphysical and intellectual energy. For son-ie teachers such energy is channeled intointense professional effort. Huberman's (1988) study identified a category ofteachers at this stage who actively diversify their classroom methods, seek outnovel practices and often look outside their own classrooms for professionalstimulation. Another group of teachers at this stage focused their efforts onseeking promotion to administrative roles or appointment to key district orstate-wide projects. Yet a third group of teachers, also identified by Sikes et al.(1985), reduced their professional commitments. Members of this group some-times experienced difficult classes and achieved poor results with their students.Building an alternative career was an option pursued by many teachers in thisgroup.

Sikes et al. (1985) estimate the fourth stage, 'reaching a professional plateau',to occur between the ages of approximately 40 and 50 to 55 years. It is atraumatic period for many; teachers at this stage are reappraising their successesin all facets of their lives. Their own sense of mortality is accentuated bycontinually being surrounded by young students and by having, as colleagues,young teachers who may be the same age as thcir own children. Responses tothis stage appear !o be of two sorts. One group of teachers stop striving forpromotion and simply enjoy teaching. These teachers may become the backboneof thc school, guardians of its traditions, and enjoy a renewed commitment toschool improvement. A second group, however, stagnates; they become bitter,cynical and unlikely to be interested in further professional growth.

Depending to a large extent on which of the two responses (discussed above)is adopted at Stage Four, teachers in the final stage, 'preparing for retirement',may behave in quite different ways. Hubcrman's (1988) study identified threedifferent patterns of behavior, each of which involved some type of contractionof professional activity and interest. One pattern of behavior, 'positive focusing',involved an interest in specializing in what one does best. Such specializationmight target a grade level, a subject or a group of students, for example.Teachers adopting this pattern, as Sikes et al. (1985) also foond, arc concernedcentrally with pupil learning, their most compatible peers, and an increasingpursuit of outside interests. A second pattern of behavior, 'defensive focusing',has similar features to the first but a less optimistic and generous attitude towardtheir past experiences with change, their students, and their colleagues. Finally,Huberman (1988) labels a third pattern of practice 'disenchantment'. Peopleadopting this pattern are quite bitter about past experiences with change and theadministrators associated with them; they are tired and may become a source offrustration for younger staff.

Implications: Guidelines for School-Leaders in Fostering TeacherDevelopment

An explicit, defensible conception of teacher development provides a foundationupon which school-leaders can formulate their own approach to teacher develop-



Teacher Development. A Central Problem for Leaders

ment. In this section, four broad guidelines for building this approach are sug-gested. These guidelines stress the importance of attending, in parallel, to allthree dimensions of teacher development and creating school cultures and struc-tures hospitable to such development (a matter pursued in more depth in Chapter9). Based on assumptions about teachers, as adult learners actively involved inbringing meaning to their work, the guidelines stress the importance of under-standing teachers' own views of their world. Finally, the guidelines argue that themost helpful teacher-development strategies available to principals are to befound among their normal responses to their work environment,

Guideline One: Treat the Teacher as a Whole Person

As Table 8.2 indicates, growth in professional expertise consists of teachersexpanding their instructional repertoires, responding more flexibly to classroomcircumstances, and taking responsibility for the welfare and growth of not onlystudents but also their professional colleagues. While the acquisition of know-ledge and skill concerning instruction, as well as other educational matters, is anobviously necessary condition for such growth, it is not sufficient. That is, thepractice of instructional flexibility depends on at least being able to weigh avariety of instructional alternatives. Many instructional strategies also require theteacher to relinquish exclusive control over classroom activities and to truststudents to be task-oriented on their own or in groups. This suggests thatprerequisite to acquiring instructional expertise (Stage Four of the professionalexpertise dimension) is growth to at least the middle stages of the psychologicaldevelopment dimension, as depicted in Table 8.1. Similarly, practices associatedwith professional expertise at Stages Five and Six appear to depend on: the abilityto synthesize alternatives, mutuality in interpersonal relations, the ability to copewith conflicting needs and duties, and other attributes of functioning at thehighest level of psychological development. Indeed, failure to attend to the inter-dependence of professional expertise and stages of psychological developmentoffers an additional explanation for lack of application, in the classroom, of skillsacquired through training. To this point, the most compelling explanation forthis 'transfer' problem has been limited to the unique and often overwhelmingdemands placed on teachers' application of newly acquired skills by their par-ticular classroom contexts Uoyce and Showers, 1980).

Typically, staff-development efforts (whether by principals or others) do notacknowledge the interdependence of psychological and professional develop-ment. While this may be due to ignorance or oversight, in sonic cases, it mayalso be due to the commonly held view that psychological development iscompleted by adulthood. That such a view is unwarranted is clear, however,from the evidence reported by Harvey (1970) that a large proportion of teachersin his sample were at the lowest level of conceptual development; Oja's (1981)review of similar evidence suggests that teachers typically stabilize in the middlestages of psychological development.

So far, our attention has been limited to die relationship between psycholo-gical development and the development of expertise. What of career-cycledevelopment? The development of professional expertise seems to have an im-portant relationship with such development. There is. for example, an obvious link


Developing Expert Leadership for Future Schools

between the challenges facing a teacher in the first three stages of his or her careercycle and the expertise to be acquired in the first four stages of development ofprofessional expertise. Indeed, interventions designed to promote the develop-ment of such professional expertise seem likely to ensure positive career-cycledevelopment. School-leaders have an opportunity to prevent painful beginnings;they are preventable through such interventions as realistic classroom assignmentin combination with ongoing assistance in the development of classroom-management skills, provision of a supportive mentor close at hand and theavoidance of heavy-handed supervision practices. On the other hand, failure toprovide opportunities for the development of professional expertise may welllead to professional disaffection when teachers are seeking new challenges andhave new concerns. Providing opportunities to master an expanded, flexiblerepertoire of instruction techniques seems an effective way of ensuring thatteachers experience a sense of professional self-fulfillment during this third stagein their career cycle.

A direct relationship appears also to exist between the career-cycle stage'reaching a professional plateau' and Stages Five and Six in the development ofprofessional expertise. A significant part of the explanation for teachers perceiv-ing themselves to be at a plateau is the failure, in many schools and schoolsystems, to permit teachers greater scopc to know and relate to multiple class-rooms, to see and work with other teachers and their classrooms. Such chal-lenges respond to the teacher's readiness to accept more responsibility and allowthe school and school system to benefit from their accumulated expertise.Teachers who have experienced such challenges seem likely to enter their finalcareer-cycle stage either still in an expansionary frame of mind or at least as'positive focusers', to use Huberman's (1988) term.

In brief, then, school-leaders should be sensitive to all three developmentdimensions and seek to help teachers develop these dimensions in a parallel,interdependent fashion.

Guideline Two: Establish a School Culture Based on Norms (if TechnicalCollaboration and Prqessional Inquiry

While teachers often appear to stabilize in the middle stages of psychologicaldevelopment, the reason is inadequate stimulation not some innate shortcomingof teachers (Sprinthall and Theis-Sprinthall, 1983). Such is the case with profes-sional expertise, as well. Evidence suggests that the typical school culture and itsorganizational structures may be responsible, in part, for stifling teacher develop-ment (for this discussion, culture includes the underlying assumption, norms,beliefs, and values that guide behavior among professionals in the school'.

Typical school cultures are characterized by informal norms of autonomyand isolation for teachers (Lortie, 1973), as well as entrenched routthes andregularities (Sarason, 1971, Leiberman and Miller, 1986). Indeed, some aspects ofthese cultures have been dubbed sacred (Corbett, Firestone and Rossman, 1987)and, as a result, highly resistant to change. Teachers' individual, personal beliefsabout the needs of students are far stronger influences on their classroom prac-tices than other potential influences such as the views of their peers or principalsor prescriptions contained in curriculum policies (Leithwood, Ross and Mont-



Teacher Development: A Central Problem for Leaders

gomery, 1982). Such autonomy and isolation limit the stimulation for furtherdevelopment to what is possible through private and unguided reflections onwhat one reads and experiences outside the classroom, and one's own informalclassroom experiments. It is unlikely that such stimulation will create the sort ofdissonance or challenge to one's ways of thinking that appears necessary to fostermovement from one stage of psychological development to another. Nor wouldsuch stimulation provide the conditions outlined, for example, by Joyce andShowers (1980) for the successful application of new instructional skills to one'sclassroom. Little (1982, 1985), on the other hand, found that staff-developmentefforts were most successful where a norm of collegiality and experimentationexisted.

School-leaders' efforts to foster teacher development seem most likely to besuccessful within a school culture in which teachers are encouraged to conscious-ly reflect on their own practices (Oberg and Field, 1986), to share ideas abouttheir instruction, and try out ncw techniques in the classroom. School-leadersneed to develop norms of reflection through the substance of their own com-munication with teachers and the examples of their own teaching; they also need

to take specific actions to foster norms of collaboration. As Rosenholtz (1989)points out, 'Norms of collaboration don't simply just happen. They do notspring spontaneously out of teachers' mutual respect and concern for each other'.Rosenholtz identthes four conditions which influence the extent to which teachersare likely to engage technical collaboration: teachers' certainty about their ownInstructional competence and hence self-esteem; shared teaching goals; involve-ment in the school's technical decisions; and team teaching opportunities whichcreate the need to plan and carry out instruction with colleagues.

This guideline suggests, in sum, that school-leaders look below the surface

features of their schools at how teachers are treated, what beliefs, norms and

values they share and redesign their schools as learning environments forteachers as well as for students. Chapter 9 provides more detailed advice on howthis might he done.

Guideline Three: Careliilly DiaRnose the Starting PointsprTeacher Development

Teachers are not passive recipients of school-leaders' strategies `to develop them'.Adopting the view of contemporary cognitive psychology (e.g., Schuell, 1986,Calfee. 1981), particularly as it has been applied to research on teacher thinking(e.g., Clark and Peterson, 1986), teachers actively strive to accomplish implicit orexplicit goals they hold to be personally important in their work. For example,when teachers judge a new form of instruction, of which they become aware, aspotentially helpful in accomplishing such goals, they make active attempts tounderstand and assess that new form of instruction. The primary resources usedby teachers to develop such understanding arc what they already know (ascontained in their long-term memory). Understanding develops: as matches aremade between the new form of instruction and what they already know (e.g.,'Oh! "Direct Instruction' means the traditional instruction I was taught inteachers' college.'); and existing knowledge structures are modified to accom-modate novel aspects of the new form of instruction (e.g., 'Ah! "Cooperative



Developing Expert Leadership for Future Schools

Learning" just means grouping with different rules that 1 have used.'); and/or aslinks are established among previously unconnected pieces of information in theteacher's memory (e.g., 'I think "mastery learning" is a combination of what Icall behavioral objectives, criterion-referenced testing and remedial teaching.').These brief examples make clear that successful ways of fostering developmentbuild on a careful diagnosis of the relevant knowledge already possessed byteachers. Such strategies will assist teachers to identify such aspects of what theyalready know and to use that knowledge as an instrument for giving meaning tonew practices which they may wish to understand better and use.

The formal mechanism most obviously available to school-leaders for carry-ing out this diagnosis is teacher evaluation. Virtually all formal school-leaders(principals) spend considerable time doing it. Nevertheless, such evaluation, as itis normally practiced, rarely results in useful diagnostic information and generallyappears to have little influence on teacher development (Lawton et al., 1986).Recent research has provided some useful clues for how principals can redesigntheir approaches to teacher evaluation, so as to be a more effective 'needs-assessment front end' for teacher development (e.g., Stiggins and Duke, 1988).For example, such evaluation needs to be based on criteria or goals that bothprincipals and teachers agree are relevant to teacher development. Multiple formsof data should be collected as a more powerful means of accurately reflectingteachers' practices and needs; regular observation of classroom practice withconsiderable time in the classroom is an important part of such data collection.The formality, frequency, and length of evaluation should be adapted to indi-vidual teachers' characteristics and needs. Rosenholtz (1989) found that teacherevaluation with features such as these was one of four organizational factorscontributing directly to teacher-learning opportunities in the school (the otherfactors were school goal setting, shared values, and collaboration).

In sum, this guideline reminds school-leaders that development is an in-cremental process which builds on teachers' existing stock of attitudes, know-ledge, and skill; they are at the same time the objects o., and instruments for,development.

Guideline Four: Recast Routine Administrative Activities into PowerfulTeacher-Development Strategies

Many school-leaders, in principal or vice-principal roles, remain skeptical aboutthe contribution which they can make to teacher development. Their skepticismis rooted in the belief that useful development strategies would include, forexample, detailed planning of in-service programs, creation of large amounts ofteacher release time for participation in such programs, and perhaps actingthey:selves as in-service instructors. It is not usually the lack of know-how thatcauses school-leaders the most despair in the face of such strategies. Rather, thedespair is caused most directly by the lack of congruence between the demandssuch formal strategies place on school-leaders' work and the real demands of thatwork. The point of this guideline is to argue that such a view of how teacherdevelopment can be fostered is essentially rrisguided. As Pfieffer suggests:'Teachers don't need Superman - Clark Kent or Lois Lane will do just fine'(1986, p. 4). Indeed, the more informal strategies available to school-leaders in



Teacher Development: A Central Problem for Leaders

their normal responses to the cicinands of the job can be much more effective infostering teacher development than such formal, hard-to-implement strategies.Expert school-leaders have learned this lesson well.

What are the 'real' demands faced by school-leaders in their work? Aspointed out several times in the book already, we know that principals' activitiesare typically characterized by brevity, fragmentation, and variety (Bradeson,1986, Davies, 1987, Gaily, 1986, Martin and Willower, 1981, Willower andKmetz, 1982). Rarely, it seems, do principals spend more than ten minutes at atime on a single task; and they make about 150 different decisions in the course ofan average day. Communication of one sort or other the primary nature ofmost principals' activities; almost three-quarters of such activities are interperson-al and take place with only one other person, over hal- involving face-to-facecontact. Formal school-leaders' work environments also require high levels ofspontaneity; the largest single expenditure of a formal school-leaders' time isreported to be unanticipated meetings.

While most formal leaders experience the demands just described, recentresearch suggests at least one compelling source of difference in the responses ofexpert as compared with non-expert school-leaders (Lenhwood and Montgom-ery, 1986). What is different is the amount of consistency that principals are ableto bring to their activities and decisions. Non-experts approach these activitiesand decisions in a relatively piecemeal fashion: for example, decisions aboutbudget, discipline, timetabling, reporting, and staffing all may be based ondifferent criteria. As a consequence, the overall effects of these decisions maywork at cross purposes.

In contrast, experts base their decisions and actions on a relatively consistentset of criteria: they 'can articulate direct and remote links between their actionsand the instructional system' iBossert, 1988, p. 348). As a result, the effects of themany, seemingly trivial, unrelated, and often unanticipated decisions made bythese experts eventually add up to something; their impact accumulates in a waythat consistently fosters school improvement. And what is the glue that holdstogether the myriad decisions of expert school-leaders? It is the goals which theyand their staffs have developed for their schools and a sense of what their schoolsneed to look like and to do in order to accomplish those goals. Such a clear,detailed vision (incorporating a conception of teacher development) and its sys-tematic use on a day-to-day basis appears to be absent among non-expert school-leaders (Stevens, 1986).

This opportunistic but clearly directed approach by expert school-leadersto their work as a whole manifests itself in the strategies they use for teacherdevelopment. Such school-leaders do not attempt to deny the fragmented, inter-personal, and spontaneous demands of the job (as would be required by a formal,in-service training approach to teacher development); on the contrary, they adaptand build on strategies that are part of their normal responses to their workdemands. McEvoy's (1987) results illustrate, more specifically, the typcs ofsubtle, sometimes opportunistic teacher-development strategies used by experts.In this study, twelve elementary and intermediate principals were observed usingsix strategies: informing teachers of professional opportunities; disseminatingprofessional and curriculum materials to teachers with personal follow-up anddiscussion; focusing teachers' attention, through meetings and informal contacts,on a specific theme in order to expand the concepts and practices teachers



Developing Expert Leadership for Future Schools

considered; soliciting teachers' opinions about their own classroom activities, aswell as school and classroom issues, thereby contributing to a sense of collegialityamong staff; encouraging teachers' experimenting with innovative practices andsupporting their efforts; and recognizing, sometimes publicly, the achievementsof individual teachers.

Examples of other ways of fostering teacher development used by expertsare provided in Leithwood and Montgomery's literature reviews (1982, 1986).These strategies include: working alongside individual teachers in their classes toresolve problems or implement changes; helping staff gain access to outsideresources; and helping teachers arrange to observe other teachers in other schools.Even relatively 'impersonal' strategies normally available to school-leaders maybe designed in such a way as to foster teacher development. Hannay and Chism(1988), for example, found that teacher transfers could become an effective meansfor fostering such development when the transfer promptcd teachers to re-examine their practices.

Wilson and Firestone (1)87, p. 20) refer to most of the strategies that havebeen mentioned as 'linkage strategies' and show how school-leaders' fosteringof both bureaucratic and cultural linkages can lead to teacher development.Bureaucratic linkages (such as creating more free time for teachers) can affecthow teachers interact with each other. Cultural linkages (such as introducingmore consistency into school communications) work on the consciousness ofteachers 'by clarifying what they do and defining their commitment to the task'.

Effective school-leaders, in sum, use the energy and momentum creatednaturally by the demands of their work for purposes of teacher development.They have redefined the problem as the solution.


Gideonese (1988, p. 65) has suggested that the teaching profession is 'undergoingrevolutionary transformation', although many of us arc too close to see it, assuch. Such change appears to begin from a perception of teaching as a routine jobconducted with craft-like knowledge, in isolation from other adults, in a hier-archical status structure. The new perception of teachthg, in contrast, views it as anon-routine activity drawing on a reliable body of technical knowledge andconducted in collaboration with other professional colleagues. Awareness of thisshift has been fostered by recent, effective schools' research and proposals in-cluded among second-wave reforms in the US (Bacharach, 1988).

Nevertheless, we need to devote much more attention to how this newlyperceived image of the teacher can be realized. This chapter has outlined plausiblestages through which teachers are likely to grow as they acquire the attributesassociated with a collaborative, professional image of the role. Sonic generalguidelines school-leaders might follow in fostering such teacher growth, havealso been proposed. These guidelines only touch the surface of a problem whichrequires much further thought, however the implications, for the role of theformal school-leader of an image of teaching as a collaborative, professionalenterprise. Only when we have clearly conceptualized coherent images of bothteacher and school-leader roles and how they develop, will we realize the com-bined contribution toward student learning of those in both roles. Much of the



Teacher Development: A Central Problem for Leaders

knowledge required for this task is already in hand. While more knowledge willbe useful, using what we already know is a crucial and immediate challenge. Oneof the most promising contexts for using what we know is in the preparation ofleaders for future schools. Such preparation, among other things, would providethem with the capacity to follow the four guidelines for teacher developmentdiscussed in this chapter.


1 This chapter ts adapted from Leithwood (199M.

Chapter 9

Collaborative School Cultures:A Key Part of the Solution

Teacher development, we argued in Chapter 8, is a key problem to be addressedby leaders of future schools. When they arc successful in fostering such develop-ment they have provided transformational leadership to their schools. In additionto providing a description of what such development might look like, Chapter 8also offered several guidelines useful to school-leaders in fostering teacher de-velopment. One of these guidelines was to help create a morc collaborativeculture among the teaching staff in the school. We offered little concrete advice inChapter 8, however, about how school-leaders might do this.

This chapter explores in greater depth what developing a more collaborativeculture means and why it is a key part of the solution to the teacher-developmentproblem. And, of greatest practical consequence, this chapter identifies specificstrategies school-leaders might use to encourage greater staff collaboration.

Interest in school culture is widespread at present, and for good reasonCurrent reform initiatives in Canada, the US, Australia, and a number of otherdeveloped countries are calling for the restructuring of schools. One centraldimension of such restructuring is the empowerment of teachers within a schoolculture that is both shared and technical (Gideonese, 1988). Such cultures notonly foster the types of outcomcs for students that are valued by educationalreformers but stimulate continuous professional growth among teachers, as well(Rosenholtz, 1989, Little, 1982).

While evidence about the positive effects of shared, technical, school culturesis growing rapidly, very little is known about how they develop (Joyce. Bennetand Rolheiser-Bennet, 1990, Fullan, 1990). Furthermore, there has been verylittle empirical research inquiring directly into what school-leaders might do toassist such development, although much evidence has accumulated in support ofthe school-leader as a crucial agent in realizing a number of other importantreform objectives (sec Chapter 2). To fill this void, wc initiated the study onwhich this chapter is based.

Framework and Methods

Extending the definition of school culture begun in Chapter 8, it is usefullydefined as:


1 "'I s


Collaborative School Cultures: A Key Part of the Solution

a system of ordinary, taken-for-granted meanings and symbols withboth explicit and implicit content that is, deliberately and non-deliberately, learned and shared among members of a naturally boundedsocial group. (Erickson, 1987, p. 12)

A school's culture consists of meanings shared by those inhabiting the school.Schools may include several subcultures as well: for example, a student sub-culture arid a professional staff subculture, the focus of this study.

Attention to school culture, as part of school reform, is driven by evidencethat traditional school cultures, based on norms of autonomy and isolation, createa work context in which realizing the central aspirations of school reform ishighly unlikely. Such norms begin to develop early in a teacher's career, perhapsduring teacher training (Su, 1990). Isolated cultures have been described byFeirnan-Nemser and Floden (1986) in terms of norms of interaction with stu-dents, teachers, administrators, and parents. Norms of authority and disciplinealong with a competing need for close personal bonds characterize teachers'interactions with students. Typical norms act to isolate teachers from asking theirpeers for, or offering to their peers, professional advice. Tearhers, it has beensaid, have peers but no colleagues. School administrators are valued by teacherswhen they act as buffers from outside pressures and maintain school discipline,but not if they interfere in daily routines or instructional decisions. Parcnts arevalued as supports for the teacher's plans and practices but are not expected to'interfere' in those plans. As a whole, these traditional norms of interaction createa highly autonomous professional culture, one that is clearly adaptive under someconditions, such as: traditional expectations for student outcomes in some typesof schools; school-leaders unable to provide instructional leadership; little publicinterest in accountability and modest expectations for the contribution of schoolsto society with few external pressures for change; prevailing images of teachingas craft (or art) based on limited technical know-how; and traditional contribu-tions by the family to the development of students.

Since most of these conditions no longer prevail in many schools, it is notsurprising to find evidence of a different teaching culture emerging (e.g., Little,1982, 1990, Nias, Southwork and Yeomans, 1982, Rosenholtz, 1989, Schneiderand Hochschild, 1988). This culture is student-centered and based on norms ofinteraction with students that are supportive and positive; while discipline ismaintained, it is obviously to serve the interests of learning readier than an endin its own right. Teachers have a shared, technical culture built on norms ofcollegiality, collaborative planning, and continuous improvement. Staff and thestudent body are cohesive and have a strong sense of community. There isreciprocity between, and among, staff and students. Administrators are expectedto offer instructional leadership and parents are considered co-partners in theeducation of students wherever possible. Such a culture appears to be adaptive toincreasingly prevalent conditions associated with calls for reform such as: newand more complex expectations for student outcomes; school-leaders able toprovide instructional leadership; high expectations by the public for its schoolsand many associated, external pressures for change; a rapidly expanding body oftechnical know-how conccrning instruction; and changing family environments.This culture is central to the 'second wave' of reform in the United States(e.g., Bacharach, 1988). Gideonese (1988) characterizes it as a 'revolutionary


Developing Expert Leadership for Future Schools

transformation' in the teaching profession, as was noted in Chapter 8, and Fullanand Connelly (1987) use it as the basis for their recommendations for reformingteacher education in Ontario.

Recently, Andy Hargreaves (1990) has identified two forms of teachingcultures in addition to those we have referred to as isolated and (truly) collabora-tive. 'Balkanized' cultures, common in secondary schools with department struc-tures, feature substantial collaboration within teaching subgroups but little or nosignificant collaboration across such groups. 'Contrived' collaboration existswhere professional interaction is mandated (perhaps by a school administrator)but where the norms of the participants would not support such interaction if themandate were removed.

Hargreaves and Wignall (1989) have also provided compelling reasons why,even within the context of school cultures which strongly support collaboration,there arc legitimate reasons for continuing to value teachers' individuality. Anethic of care, posit Hargreaves and Wignall, drives many teachers to spend asinuch time as possible in contact with their students. Further, while the benefitsof collegiality may include spurs to creativity and effective professional problem-solving, solitude may sometimes offer the same advantages for some peopleAmong the goals for cultural change, then, would seem to be: the removal of'administrative or other situational constraints' (Hargreaves and Wignall, 1989,p. 15) to collegial work; the creation of norms of collegiality which neverthelessacknowledge the value of individual, autonomous work on some matters; andthe development of forms of collegial work which maximize the potential ofshared problem-solving.

The study on which this chapter is based (Lcithwood and Jantzi, 1991)addressed three questions. The first question concerned the reasons for variationin degrees to which schools achieved collaborative cultures when they set outwith that goal. Because changes in culture seem likely to occur in the course ofpursuing other goals, as well, this study was part of a more comprehensiveanalysis of school-improvement processes. Therefore, the second research ques-tion inquired about the significance of the larger set of improvement processes inwhich people engaged for the purpose of developing collaborative culturesFinally, we asked about the strategies used by school administrators to developmore collaborative school cultures. The design of the study involved focusing onschools which were known, in advance, to have achieved collaborative cultures,data were collected in an effort to determine how such cultures developed andwhat role was played by school-leaders. Twelve schools (nine elementary, threesecondary) were selected for the study, all of which had been involved in seriousimprovement eftbrts for a minimum of three years.

Schools were selected for the study through a two-step process. In the firststep, school-district adininistrators, Ministry of Education officials, and Facultyof Education professors were asked to nominate schools in their jurisdictions thathad experienced significant improvement over the past several years and werenow considered to be exemplary schools. As a second step, we developed andadministered a brief questionnaire that askcd staffs in each nominated school toindicate the extent of the change that had taken place in their schools, key factorscontributing to the change, the extent of planning involved, and the nature of theimprovement that has resulted. to date. Everyone who nominated a school was



Collaboratwe School Cultures A Key Part of the Solution

also asked to complete the same screening questionnaire as the relevant districtadministrators. Data about each school were collected from an average of elevenpeople, a total of 133 people.

The procedure used to analyze the interview data was adapted from thework of Miles and Huberman (1984) on qualitative analysis. Following a series ofintermediate data-analysis steps (described in detail by Leithwood and Jantzi,1991) a 'causal network' was produced for each school. These networks identifiedthe variables involved in school improvement in each case and the relationshipamong these variables. Based on these analyses, we now address the threequestions of interest in this chapter.

What Accounts for Variation in the Degree to which SchoolsAchieve Collaborative Cultures?

My colleagues are really good. When I was on the grade 1 team andteaching grade 1 for the first time the team took me under their wing;for example, they would leave work in my mail box. They're nothoarders. In our team planning we all took a share; we all pulled ourweIght .... There's more sharing and communication among teachersabout students, their needs, progress and problems. More teachers arcaware of student problems and styles that will make a difference.

This teachei's remarks lent weight to the claim that, within her school, there wasa collaborative or 'shared, technical' staff culture. Similar remarks can be found inteacher transcripts for all case schools. However, since this study was concernedto discover what school-leaders did to foster such cultures, more systematicassurance that collaborative cultures had been achieved to a substantial degreewas necessary.

To provide this assurance, we adopted six indicators of collaboration fromthe work of Judith Little, 1982, p. 331). Her research suggested that in excep-tionally effective schools with shared technical cultures:

teachers engage in frequent, continuous and increasingly concrete andprecise talk about teaching practicesteachers arc frequently observed, and provided with useful (if potentiallyfrightening) cntiques of their teachingteachers plan, design, research, evaluate, and prepare teaching materialstogetherteza hers teach each other the practice of teaching

These and other features of such schools. Little found, could best be explained bytwo prevailing norms shared by staff:

a norm of collegiality (the meaning of which is nicely illustrated in thequotation beginning this section)a norm of continuous improvement: staff are motivated to grow profes-sionally and to contribute where possible to school improvement

1 5 0131


Developing Expert Leadership for Future Schools

Given these four specific and two more general indicators of collaboration, acontent analysis was carried out of the transcribed interviews conducted with theprincipal and all teachers in each of the twelve case schools. Its purpose was toestimate the degree of collaboration within each school. Results suggested thatthe cultures of the twelve schools, as a whole, were characterized by relativelyhigh degrees of collaboration. Across all criteria, an average of approximately70 percent of staff provided evidence of collaboration. This average rises to80 percent if the observation criterion is not included. Teachers provided littleevidence of frequent classroom observation and feedback (Mean = 16 percentconsistent with evidence reported earlier by Little, 1982). Observation and espe-cially critical feedback may violate what Corbett, Firestone and Rossman (1987;refer to as a `sacred' norm of teaching. This non-evaluative norm, with respect toone's peers, is often reinforced in codes of ethics promoted by teachers' profes-sional associations. The greatest percentage of teachers provided evidence ofcontinuous, practical, concrete talk about teaching practice (M = 95 percent).

In spite of the high levels of collaboration evident in the schools as a whole,there was significant variation across the schools. To better understand rcasonsfor variation, the three schools with the lowest levels of collaboration werecompared with the three schools having the highest levels. This comparisonsuggested two reasons for variation in levels of collaboration. First, in the threelowest scoring schools, there was little indication, in the initial stages of theimprovement projects, of strong motivation among staff members to participate.Their commitment to the project emerged in all cases but much later in theprocess. In contrast, staff motivation to participate was strong from the begin-ning in two of the three high-scoring schools; in the third, the principal quitequickly replaced staff who were not keen to participate with staff who were.Early and sustained motivation of teachers to engage in school improvement,therefore, may be a crucial determinant of at least the speed with which col-laborative cultures develop. Actions by school-leaders to foster early enthusiasmamong teachers for the school-improvement effort may contribute significantlyto their predispositions toward collaboration.

A second reason for differences in levels of collaboration across schools maybe found in the variable we label 'goal clarification'. The three schools whichachieved the most collaborative cultures arrived at a sct of clear, shared goals forschool improvement in the context of substantial staff cohesiveness and/or col-laborative decision-making. Goal clarification in two of thc other three schools,in contrast, was stiniuLted direitly by the school-leader or a leadership teamprior to the development of much staff cohesiveness or collaborative decision-making. In the third school, goal clarification did not occur (we comment inmore detail later on the importance of goal clarification).

The nature of collaboration varied across the schools in different ways aswell. For example, in one school there was much evidence of joint work but thiswas mostly work in pairs and almost all of it involved the principal as onemember of the pair. While this form of collaboration seems to have been usefulin meeting the immediate school-improvement goals. it leaves the school culturcespecially vulnerable, should there be a change of school-leader. Collaboration inthe three secondary schools also seemed different than in elementary schools. Inthese schools, joint work was usually within departments; they were 'Balkanized'in some respe(ts (Hargreaves. 1990) Nevertheless, in-service work was frequent-


1 5 1.

Collaboratwe School Cultures. A Key Part of the Solution

ly carried out across departments as was talk about teaching. Overall, the normsof collegiality and continuous improvement were as evident in secondary as inelementary schools.

In sum, we interpret these data as evidence of school cultures showingrelatively extensive collaboration, although variation among schools existed for

several reasons.

How Significant is the Larger Context for Cultural Change?

The (School Improvement Project) gave a more integrated and unitedfocus on what every individual teacher would try to do, things thatnormally wouldn't be spelled out. It helped unify the school's instruc-tional focus and the things we wanted to do through learning wereimportant for everyone ... we had support all along the way, we justtalked to each other for support (Teacher).

The larger context within which each of the school-leaders in our sample oftwelve schools worked to develop more collaborative cultures was a school-improvement project. The specific purpose of these projects was determined byeach school, sometimes in conjunction with district staff, and varied widely.Some projects focused on staff development, some on the implementation of newcurricula, others attempted to use library resources more effectively and increaseteachers' repertoire of instructional strategies. One secondary school was in-volved in a massive effort to individualize instruction, an initiative that touchedon virtually every aspect of the school's program, administration, physical organ-ization, and culture. Most improvement initiatives were much more modestthan this.

To clarify the larger context in which school-leaders worked to develop ainure collaborative culture, we have selected one of the twelve case studies ofschool improvement as an illustration. This was a suburban, K-8 (kindergartento grade 8) school with an ethnically and socially diverse school population ofabout 400 and a teaching staff of twenty-one members. At the beginning of theproject (early 1987) the principal had been at the school for three years. She wasencouraged to develop the project, initially, by a senior administrator, after shehad expressed some unease about instruction in the school. At the outset,teachers also expressed concern about lack of staff cohesiveness. The focus forschool improvement became mastering the use of inquiry teaching methods andusing them appropriately in the classroom.

Figure 9.1 is a causal network depicting the twenty-eight variables involved

in the school-improvement processes that took place over three years in thisschool. Complete definitions of these twenty-eight variables, as well as otherswhich appeared in the twelve cases as a whole, arc available from the authors.For present purposes, variables referenced directly in the text are briefly definedwhen they are discussed. Figure 9.1 also depicts the relationships among vari-ables. Based on the frequency with which variables were mentioned by the tenstaff interviewed for this case, an estimate is provided of whether the influence ofthe variable was low, moderate, or high. In cases where the relationship isnegative, a '(-)' appears next to the line joining the variables. Because the present


Figure 9 1 School-improvement processes Causal network for one school












3 pressure

L 4






7.144411k iLnef al udeenr peh


L A[Staffmertia



/ 4)-



M 12







14 /Goalclarification


Physical andprogramadiustments


















t 27






Main Paths Variable Influence

* 1 H = high

2 M = moderate

3 L = low 1 3

Collaboratwe School Cultures A Key Part of the Solution

study is focused on the development of collaborative cultures, it is useful to notethat five of these twenty-eight variables constitute aspects of such a culture:collegial support (15), collaborative decision-making (20), staff cohesiveness (18),strengthened relationships (25), and modification of school culture (28).

A brief narrative will help the reader understand the processes depicted inFigure 9.1. There were four antecedent variables, of which three had differingbut overlapping paths throug: ; the mediating variables to the outcomes. District-level endorsement (1) of the Ministry School-Improvement Project (SIP) and itsinvitation to this school was a factor in school-administration endorsement (2) as

was the principal's perception of staff inertia (4), which had led the schooladministration to undertake a needs/capacity assessment (3) and to create pressure

for change (7).The first path begins with board endorsement (1) of the school being

designated a SIP school. This endorsement led to board support (9) as a mediat-ing variable in the change process. This support was manifested in available

resources (10) as board consultants provided some support, and in monetarysupport (13), particularly for release time. Both the monetary support and theresources contributed to staff development (17). During the process of theirstaff-development activities, teachers had made revisions to the board curriculumdocument that resulted in some external recognition (19). This acknowledgmentof their work was a factor in teacher's altered attitudes (22) to greater confidence

in their professional abilities. The staff-development variable was related to other

outcome variables in patterns that will be described below.The second and most significant path through the causal network begins

with school-administration endorsement (2) interacting with the needs/capacityassessment (3) to initiate the school-improvement effort. This endorsement wasreflected in the mediating variable, school-administration support (6), which was

reinforced by both Ministry (5) and board support (9). School-administrationsupport contributed directly to staff development (17) as well as contributingindirectly through the principal's efforts to obtain whatever limited resources (10)

that were available. School-administration support (6) also resulted in thedevelopment of leadership influence (11) with the formation of the school-improvement team (SIT) which connected with the third path through collabora-tive decision-making (2(J). Leadership influence from the School ImprovementTeam (SIT) led to increased goal clarification (14). which helped to alleviateanxiety and stress (12), as teachers developed greater certainty about what theycould do; the clarification process also contributed to staff development (17). Anexternal expert (16) was also crucial in develz.ving greater clarity about goals (14)

and in staff development (17). As an outcome of staff development, attitudesaltered (22) and knowledge and skills increased (23), leading to changed behaviorand actions (26) as well as professional growth (27). These outcomes, in turn,contributed to modification of the school culture (28). Staff development contri-buted to physical and program adjustments (21) to accommodate the inquiryteaching model. These adjustments led to policy changes (24) and strengthenedrelationships (25) with staff engaged in joint planning. These outcomes alsoaffected professional growth (27), behavior and action (26), as well as schoolculture (28).

The tinal path through the network begins with the antecedent variable, staff

inertia (4), which influenced the principal to initiate a needs/capacity assessment



Developing Expert Leadership for Future Schools

(3), and led to school-adminiszration pressure (7); staff were given a choice ofmaking a three-to-five-year commitment to the SIP or transferring to anotherschool, and new staff were hired only if they supported the SIP. This pressureresulted in a high level of user commitment (8) that contributed to collegialsupport (15), since teachers were willing to help others so the SIP would be asuccess. Some of the collegial support was the result of school-administrationpressure (7) to form planning partners. The pressure to be active in the SIP alsoresulted in some anxiety and stress (12), as teachers were uncertain how toproceed. The anxiety was partially offset by increased collegial support (15). Asstaff supported each other in their SIP activities, there was an increase in staffcohesiveness (18), with new and previous staff pulling together as a team.Cohesiveness had a reciprocal relationship with collaborative decision-making(20), which was supported by leadership influence (11) from the SIT. Greatercollaboration in decision-making was part of the process for developing clarity ofgoals (14), and contributed to staff development (17). Staff development, in turn,enhanced collaborative decision-making among staff colleagues and within class-rooms, as staff became more proficient in use of the inquiry teaching method.Collaborative decision-making (20) along with staff development (17) contri-buted to physical and program adjustments (21) to support the new teachingstrategy. In this interaction with the second path, policies were changed (24) andrelationships strengthened (25), as noted above. Collaboration also resulteddirectly in strengthened relationships among staff, among students, and betweenstaff and students. Change in relationships was reflected in behavior and actionchanges (26), and teachers' professional growth (27), as well as contributingdirectly to school culture.

Figure 9.1 depicts a comprehensive set of school-improvement processessimilar, in many respects, to processes found in the other eleven case studies.What significance does this larger context have for the development of collabora-tive cultures? First, It is important simply to recognize that while 'restructuringschools' (in this case their cultures) is a worthwhile reform objective, it isunlikely that people will be motivated to pursue it as an end in itself. Consistentwith the results of other research (e.g., Huberman and Miles, 1982), our casestudies suggested that most people were motivated to change by goals muchmore directly cc,ncerned with curriculum and instruction.

Second, the larger context for school improvement, usually focused on'educationally compelling' changes, potentially neutralizes the primary disincen-tive for increased collaboration with colleagues: the significant additional costs ofcollaboration, particularly in the initial stages. Increased costs are a function ofthe time required for group deliberation, as distinct from individual decision-making. This time may be added onto the normal work ime of staff or requirethe direct outlay of money to free people from duties during their normal worktime. The incentive for collaboration is the perception that these costs are at leastbalanced by significant benefits. Such perceptions depend on collaboration beingviewed, eventually at least, as a powerful means for coping with problemsattendant upon implementing 'educationally compelling' changes.

A third way in which the larger context for school improvement has signi-ficance for developing collaborative cultures is the multiple and diverse opportu-nities it affords to reinforce the benefits of collaboration. This can be illustratedwith reference to Figure 9.1. for example, by examining the relationships with

I 36


Collaboratwe School Cultures A Key Part of the Solution

collegial support (15). This variable emerges as a direct response to pressure from

the school administration (7) to begin to implement inquiry teaching strategies,

and a parallel commitment to that goal by many teachers (8). The development

of collegial support, in this case, initiates a process which eventually increases

staff cohesion (18) and results in collaborative decision-making. (20). But the

perception that collegial support is valuable is usefully reinforced by its ability

also to help teachers cope with an unintended by-product of school-administration pressure: increased anxiety and stress (12).

Finally, as Figure 9.1 also illustrates, at least understanding the larger context

for school improvement allows us to appreciate how significant a force school-

leaders can be in the development of more collaborative cultures. For example,

the principal's support for instructional change (6) in this case study was followed

by delegation of leadership responsibilities to others and a school-improvement

team (11), and the need for staff development (17). Both of these variables

subsequently fostered a need for collaborative decision-making (20). School-

administration pressure (7) was similarly pivotal in the development of otheraspects of a collaborative culture.

What Strategies Were Used by School-Leaders to Influence School


Understanding the larger context within which a collaborative culture developsdraws attention, as we have noted, to the extent of school-leaders' potential

influence on that process. In addition, such understanding helps clarify more

specifically what school-leaders can do in exercising such influence. This matter is

addressed here in two stages. First, variables most directly linking school-leaders'

actions to collaborative cultures are discussed. Then, a more detailed analysis of

specific leadership strategies is provided.The relationship between strategies initiated by school-leaders and school

culture is neither simple nor direct (as Figure 9.1 illustrates). To better under-

stand the nature of the 'space' between the two, we developed a series of

cross-case causal fragments (Miles and Huberman, 1984) focused, however, on

only selected subsets of variables and relationships to best represent the twelve

cases, as a whole.We first examined the most direct chains of relationships between school-

leaders' actions, and one central attribute of collaborative cultures strengthened

interpersonal relationships among staff that cont-ibuted to changes in school

cultures. School-leaders' actions, for this purpose, were considered to be of only

two types: those intended to be helpful, supportive or facilitative (e.g., provision

of resources) and those intended to exert pressure for change on teachers (e.g ,

persuasion of reluctant participants).Thc space between supportive school-leader actions and strengthened staff

relationships is filled with six variables, two of which, leadership influence and

collaborative decision-making, appeared most frequently. Collegial support and

organizational adjustments, appearing moderately often, facilitate the develop-

ment of collaborative decision-making directly. Collegial support interacts with

staff development and user commitment in a fw cases. The space between

1 b7


Developing Expert Leadership for Future Schools

school-leader pressure and strengthened relationships is filled with four variables,none of which appeared in many of the cases.

These CWO causal fragments sugvsted, in sum, that:

the most direct contributor to the development of strengthened rela-tionships is for staff to be involved in collaborative decision-makingthe likelihood that staff will participate authentically in such decision-making is a function of the amount of support they perceive fromcolleagues, their commitment to accomplishing their school-improve-ment goals, and the opportunities for collaboration provided throughadjustments to the organization (e.g., time to meet, suitably structuredgroups)school-leaders have at their disposal activities which arc reasonably effec-tive in creating user commitment and suitably adjusting to the organiza-tion ... delegating power to others in the school seems likely to lead togreater collegial support and forms of staff development that assist inbuilding collegial support

In order to identify more specific leadership strategies, we relied on thetwelve causal networks as well as the content analysis of the interviews with thetwelve school-leaders reflecting on thcir own actions to influence their schools'cultures. In reporting results, we assume high levels of interdependence betweenstrategies used by school-leaders to help implement the 'educationally compell-ing' changes in their schools and strategies influencing school cultures. Justifica-tion for this assumption can be found in the twelve causal networks.

Results suggested that six broad strategies were used to influence schoolcultures. School leaders:

strengthened the school's cultureused a variety of bureaucratic mechanisms to stimulate and reinforcecultural changefostered staff developmentengaged in direct and frequent communication about cultural norms,values, and beliefsshared power and responsibility with othersused symbols to express cultural values

Each of these broad strategies manifested itself in a range of specific actions.While all school administrators in the study used most strategies, their differentchoices of methods illustrate this range reasonably well. Each of these strategieswill be examined in more detail in the remainder of this section.

Sirenvhen the Sdwol's Culture

Initially the SIT (School Improvement Team( got together to talk aboutgoals for the SIP (School lmprovetnent Project( and then we decided totake everything back to the staff to see what kind of general direction


Collaborative School Cultures A Key Part of the Solution

they wanted. We I the whole staffi brainstormed what we needed to dowith resource-based learning in the school. We broke into groups anddecided on a variety of very tight recommendations and after discussionwere able to reduce all our ideas to the six subgoals under RBL. We thentook on the commitment of working each of these through, until wewere comfortable they had been achieved (Teacher).

School cultures are typically weak (Firestone and Wilson, 1985). As such, theyremain of little consequence in bringing about significant school reform evenwhen they reflect aspects of collaboration. Firestone and Wilson claim that suchweakness is function of (a) ambiguous, excessive, poorly specified purposes; (b)the isolation of teachers from one another and from administrators; and (c) lowlevels of commitment by staff to the school's purposes. The causal networks(Figure 9.1, for example) suggested that most school-leaders, at least implicitly,recognized the need to strengthen their schools' cultures and acted in a variety ofways to ameliorate sources of cultural weakness.

All school-leaders in this study, as in the study by Deal and Peterson (199(t)

engaged in some process to clarify and prioritize a set of shared goals for theschool-improvement initiatives (variable 14 in Figure 9.1), although variationsamong schools in how this was done have been noted already. This ofteninvolved entire school staffs in a process for setting goals, initially, sometimesusing a consultant to assist. It also involved efforts throughout the life of manyprojects to block competing priorities and systematically orient new staff to thegoals for school improvement.

Rosenholtz's (1989) research has depicted a central role for shared goals inhelping to tbster a shared, technical culture. For this reason, we examined moreclosely the relationships associated with this variable in our study and found thatthe relationship between school-leaders' actions and such goals is quite direct.Establishing collaborative decision-making procedures was a moderately frequentprerequisite to goal claritication. In more than half of the case schools, suchprocedures were used to identify goals and priorities. In most schools, the powerto establish goals for school improvement was delegated to or shared withothers, usually a school-improvement team. This fostered greater participation in

the process and prevented the principal's goals from dominating the process (orfrom being seen to dominate the process). Rosenholtz (1989, p. 15), explainingsimilar relationships in her data, suggests that 'principals who involve teachers ingenerating information about the goals of teaching, in scanning and choosing thebest alternatives, grant teachers a part in constructing school reality'.

Reducing teacher isolation, a second method of strengthening school culture,was accomplished by creating opportunities for staff to influence one another(e.g., creating time for joint planning, holding staff retreats, asking staff to offerworkshops to colleagues, encouraging teachers to visit one another's classes) andwhich sometimes required interaction (e.g.. creating working committees with

specific tasks assigned).Finally, teacher commitment was stimulated quite directly and forcefully in

at least four of the cases. Teachers were given the option. after reasonable oppor-tunities to understand the school \ purposes. to stay in the school and devotethemselves to those purposes. or transfer to another school, with the principal's



Developtng Expert Leadership for Future Schools

assistance. In most cases, as well, only new teachers were hired who expresseda prior commitment to the schools' purposes, an action Rosenholtz (1989) alsofound related to the development of shared goals.

Use of Bureaucratic Mechanisms

The principal encouraged us to meet with other people at our own gradelevel. I found that one of the best ways was to get down to business anddo some planning because it gives you a different way of looking atthings. This year two of us are working together on some units. Initiallyit was time consuming and we just plodded along doing so muchtalking, but by the time we were finished wc really understood how togo about it (Teacher).

Bureaucratic mechanisms, as Firestone and Wilson (1985, p. 278) point out,'establish constraints on and opportunities for how teachers teach'. Such mechan-isms will sometimes support cultural changes by making such changes easier toaccomplish or more rewarding in which to engage. School-leaders reported usinga number of such mechanisms to foster directly implementation of school-improvement goals and to create more collaborative cultures. Theses mechanismsincluded, for example:

money (e.g., reallocating existing money for the project, finding newmoney, buying needed materials)planning and scheduling (e.g., providing time for collaborative planningduring the workday, timetabling students to allow teachers to worktogether, keeping school improvement on the forefront of meetingagendas)decision-making structures (e.g., establishing divisional and committeestructures, pairing teachers for planning)staffing procedures or, more specifically, what Sashkin and Sashkin(1990) termed 'value-based' staffing (e.g., selecting new staff based onimprovement priorities and willingness to collaborate, involving staff inhiring decisions)evaluation (e.g., progress with school improvement across school, super-vise improvement efforts in individual classrooms)

The last three mechanisms, especially teacher evaluation, Rosenholtz (1989.p. 27) found to contribute significantly to teachers' commitment to school goals.This contribution occurs, in her view, when teachers believe 'that evaluationcriteria are ... central to their work, applied frequently and capable of beinginfluenced by their own effort'.


Staff Devekpment

Our prMcipal shares ncw knowledge and is always questioning wherewe are, what our strengths and difficulties arc. She finds resources for us


Collaborative School Cultures. A Key Part of the Solution

if she doesn't have the answer. She monitors us and pulls together acrossdivisions. She gets outside help if she sees any need ... Sharing is thekey thing here. Because I was new I had extra help with planning. I tookmy planning to my colleagues and my principal and discussed where theproblems lay (Teacher).

Activities designed to improve teachers' skill, knowledge, understandings orperformance in present or future roles', (Fullan's, 1990, p. 3, definition of staffdevelopment), appeared prominently in all twelve causal networks. While staffdevelopment and increased collaboration were not linked by ally logical necessi-ty, they were linked empirically to at least some aspect of collaboration in allcases. The reasons for such a link are evident in Little's (1982, p. 339) researchwhich suggested that:

To the extent that school situations foster teachers' recourse to others'knowledge and experience, and to shared work and discussion, teachersare likely to favor some participation in staff development staff de-velopment appears to have the greatest prospects for influence whenthere is a prevailing norm of collegiality.

Staff development which acknowledges what can be learned from one'simmediate colleagues, as well as others, fosters a collaborative culture and is, inturn, nurtured by that same culture. This reciprocity is evident in the chains ofvariables that form the causal fragment between school administrators' actionsand staff development. The chains consist of eight variables, six of which havebeen discussed previously, as they bear on the development of collaborativecultures. Two variables, monetary support and other available resources, are notrelated in any way to collaboration hut are directly linked to the creation of usefulstaff development, as is supportive action by the school-leader. Furthermore,school-leaders have considerable influence on these two variables directly. Thiscausal fragment suggests. in sum, that useful staff development depends mostdirectly on the commitment of the staff to school-improvement goals. theperception of collegial supports, and the availability of monq: and other sourcesto support staff-development activities. School-leaders help create useful staffdevelopment directly, by providing the needed resources, and indirectly, bytbstenng staff commitment and a supportive collegial environment. Both of thesefindings are similar to results reported by Rosenholtz (1989) and by Little (1982.p. 334). Finally, delegating power to others (leadership influence) is a keystrategy for building the kind of collegial support that makes staff developmentrelatively meaningtial.

School-leaders reported fostering staff development in both direct and in-direct ways. They acted directly by themselves giving workshops to staff in areasof their expertise, assisting teachers in their own classrooms, attending in-servicesessions with staff, and sharing information from conferences or workshopswhich they had attended. Through such actions school-leaders modeled valuesconsidered important in the school (Sashkm and Sashkin, 1990). Less directly.school-leaders informed staff of in-service opportunitics and encouraged parti-cipation, invited 'experts' into the school to assist stafT, and sent staff to relevantconferences. They also encouraged use of board consultants and provided reading


Developing Expert Leadership for Future Schools

to staff and follow-up with discussion. Because these activities brought teachersinto contact with either the school-leader, other colleagues in the school, or otheradults outside the school, they encouraged the use of collaboration as a means forproblem-solving.

Direct and Frequent Communication

In actual planning in half day sessions I was actively involved with aplanning team working with consultants. It's important to workthrough with the teacher to understand each unique classroom settingand the problems that may arise.... I think the thing I've learned linthis process] is that a principal needs to learn as much as she can about ateacher. You need to know your staff thoroughly, listen and showpeople you truly care about them. When they realize you are ready tohelp them realize their goals, you will find a positive and favourableresponse (Principal).

Firestone and Wilson (1985) hypothesize that active communication of the cultureis an especially opportune strategy for principals because of the large proportionof their time spent in interpersonal contact. Some research estimates that up to 75percent of a principals' time is spent in such contact (Willower and Kmetz, 1982,Martin and Willower, 1981). School-leaders in our study frequently used suchwords as 'informing', 'persuading', 'directing', 'writing'. 'negotiating'. 'counsel-ing'. 'visiting', and 'discussing' to indicate the prevalence of this strategy and theimportance they attached to it in pursuing school improvement and greatercollaboration. This strategy is not clearly distinguishable from others we havediscussed since all, to this point, include an important -o..e for communication.What is different in this case (aside from staff development) is that the school-leader is the source of the communication and, as a result, controls its contentmore directly than in :he cases of previcmsly examined strategies.

Iknver and Respoinibil.ty with Others

he SIT ISchool Improvement Team) was involved m planning wherewe are going, looking to the future and getting input from staff in termsof where we think we need to change (Principal).

Central to the concept of a collaborative culture is the more etinitable distributionof power for decision-making among members of the school. Especially whenthe focus of the decision-making centers on cross-classroom and school-widematters, this will involve school administrators in at least delegating, if notgiving away, sources of power traditionally vested in their positions Withoutschool-leaders' willingness to do this and teachers willingness to ac cept thepower thus ofkred, true collaboration seems unlikely. With power, ot eourse.conies the responsibility for decisions and actions. This has the potential formaking the teacher's role in the school not only much more meaningful, but alsomore complex. Teachers. under these circumstances, are professionally more


1 6

Collaborative School Cultures A Key Part of the Solution

`empowered' a goal of many school-reform advocates at present (e.g., Maeroff,1989).

The most obvious way school-leaders in this study went about sharingpower and responsibility was through the establishment of school-improvementteams, of which they were sometimes members. These teams shared the respon-sibility for project coordination with principals and assisted principals in many ofthe strategies already mentioned. Also, these teams served as important linksbetween staff and administration, testing plans and soliciting reactions and ideasIndividual members of these teams often acted as mentors or role models fortheir colleagues; they shared expertise, tried out new ideas in their classes, andencouraged conversation about the school-improvement effort.

Use Symbols and Rituals to Express Cultural Values

Every staff meeting starts with 'good news' about thc school, somethingI've seen in the classroom or thcy tell us ... at the end of the year wealways have a celebration of sorts where we look back at achievementsand celebrate what we've achieved. Our newsletter mentions teachersand talks about their accomplishments ... Children have been involvedin things like the safety patrol, assemblies, and daily announcements. Astrong sense of community has been developed in the school because thechildren identify with their mascot (Principal).

Sashkin and Sashkin (1990), Deal and Peterson (1990), and Firestone and Wilson(1)85) suggest this strategy as a promising one through which principals caninfluence school culture. As others point out, symbols are visible expressions ofthe content of an organization's culture (Peters. 1979, Pfeffer, 1981, Shein, 1985)Manipulating such symbols and rituals, therefore, is a way to make more visiblethose aspects of the culture that school-leaders believe arc valuable.

In our study, school-leaders explicitly mentioned three ways in which theyused symbols and rituals to foster collaboration. At staff meetings and assembliesthey celebrated and publically recognized the work of staff and students whichcontributed to their school-improvement efforts. This action invited others toshare in the su :esses of their colleagues. School-leaders also wrote private notesto staff, expressing appreciation for special efforts. This action demonstrated toindividuals the value attached to some practices by the principal and the possibil-ity of recognition by an esteemed colleague. Staff were encouraged, third, toshare experiences with their colleagues, both as a source of stimulation forcolleagues and also for recognition by other adults.

Each of these ways of using symbols and rituals has the potential of contri-buting to an increase in teachers' professional self-esteem. This is a pivotalvariable, influential in building a collaborative culture, according to Rosenholtz's(1989) research, a variable which increases the likelihood that teachers will feel'safe' in revealing their work to others.

Our study suggests, in sum, that school-leaders used six strategies to in-fluence the culture of their schools and to foster greater collaboration. Theystrengthened the culture, modified bureaucratic mechanisms, and engaged in staffdevelopment In addition, school-leaders communicated frequently and directly



Developing Expert Leadership for Future Schools

with staff, shared power. and used symbols to express cultural values. A widerange of specific actions were taken by school-leaders to pursue each strategy andsome actions served multiple purposes.

Summary and Further Insights about the Process ofCulture Change

Compelling evidence suggest.; that collaborative school cultures contribute signi-ficantly to teacher development. Such cultures, which are 'shared and technical'appear to foster practices most conducive to the types of staff (and student)development which are the focus of current school-reform efforts. We alsobelieved that evidence, in support of the claim that school administrators canhave a significant impact on schools, was compelling but not available specificallyin relation to school culture. Indeed, in our own recent research with secondaryschool administrators, the relationship between their practices and their schools'cultures was extremely weak (Lawton and Leithwood, 1988). 'Why was this thecase?', we wondered. 'Was it because school culture has not been a focus of theirwork? Or have they tried to influence school culture and failed?' These questionsstimulated our loquiry into what nine elementary and three secondary principalsdid to help develop collaborative cultures in their schools. These were schoolsknown to have such cultures and also to have experienced success in theirschool-improvement efforts over at least a three-year period.

After identifying sources of variation in collaboration in each school, thestudy inquired about th, significance of the larger school-improvement contextfor the development of collaborative cultures. We found that this larger contextprovided important incentives for initiating collaboration, and continuously re-inforced the ,value of collaboration over time. We then asked what strategiesprincipals used to foster greater collaboration and were able to identify six. Theseincluded strengthenin,..; the culture, using bureaucratic mechanisms, fosteringstaff development, frequent and direct communication, sharing power and re-spornibility. and using rituals and symbols to express cultural values. Thesestrategies seem quite teachable and ought to be developed within the repertoire ofleaders of future schools.

The study provided support for the claim that school-leaders have access tostrategies which are 'transformational' in effect and, hence, assist in the develop-ment of collaborative school cultures. This means two things in our view:significant changes in staff mcmhers' individual and shared understar.:ings oftheir current purposes and practices; and an entranced capacity to solve futureprokssional problems. individually and collegially.

The transformational effect of the strategies identified in the study may beexplained by thc ways in which they alter the patterns of interaction among staff.Because meaning is socially constructed (Berger and Luckmann, 1%5), differ-epees in the patterns of interactions experienced by school staff will result indifferences in the mcanings which they associate with their work. The meaningthat individual staff members bring to their work is a product of tlw schemata(eoncepts or related ideas) they possess in relation to that work. Borko andLivingston (1)89), for example. identify he specific scripts, mental scenes and


Collaborative School Cultures: A Key Part of the Solution

propositional structures that make up the schemata which expert teachers bringto their instructional practices. Such schemata are the cognitive guides to action,growing out of the teacher's assumptions, norms, values and beliefs about, forexample, the role of teacher and the nature of effective instruction. Alteration, inthe school culture and accompanying changes in staff practices, therefore, dependon changes in the individual schemata guiding a staff member's practices. Suchalterations in cognitive structure, however, must eventually coalesce aroundschematic content that is common, in some optimum measure, across indi-viduals. This recapitulation of similar cognitive processes at different levels ofsocial structure (Mehan. 1984) is a more complex type of change to assist withthan schematic change within an individual alone. This is especially so, since eachstaff member's starting points will necessarily be different.

In a traditional, isolated, professional culture, these schemata are adapted,extended, and linked together in new ways primarily in response to students.Such social negotiation of meaning will usually take place indirectly: teachers willhave to infer the implications for their purposes and practices from students'responses. These responses rarely challenge teachers to reflect directly on theirbasic assumptions and values. Indeed, relying primarily on the negotiation ofmeaning with students, as Rosenholtz (1989) found, seems likely to encourage agradual narrowing of purpose, a reducing of the teacher's aspirations for stu-dents, and an increased weight given to practices which arc successful in manag-ing classroom behavior. Rarely, in an isolated professional culture, will teachers'assumptions, norms, values, and beliefs be challenged by significantly moreambitious visions of what is possible.

Collaborative cultures, on the other hand, potentially confront teachers witha different order of dissonance about purposes and practices to which they mustadapt their classroom schemata. The social negotiation of such meaning willoften be quite direct; it also lus thc power to challenge the teacher to reconsiderbasic assumptions and values. Unlikc students, other peers (including principalsas well as teachers) are much more likely to stimulate thc teacher to considermore ambitious purposes and non-trivial modifications in their practices, as away of achieving such purposes.

The potential effect of the six strategies, used by school administrators toinfluence school culture, which we identified in this study, can be explained fromthis theoretical perspective. Three of these strategies provide principals and otherteacher, with opportunities to clarify explicitly the preferred content of relevantschemata from their point of view: using symbols and rituals to express culturalvalues; direct and frequent communication; and staff development. More interac-nye versions of these strategies allow for the negotiation of schematic contentbetween school-leaders and teachers or among teachers. Several of thc sixstrategies appear to constrain thc range of schematic content available to theteacher, rather than to dictate its precise form. This seems tli case, for example,with the use of bureaucratic mechanisms. Finally, sharing power and responsibil-ity may provide a stimulus for developing shared meaning (effecting the strengthof the culture) without necessary reference to the meaning itself (the content ofthe culture). If school-leaders share power with those who hold similar pointsof view, however, this strategy is more prescriptive of the schematic contentpwcd hy school-leaders.

1 6 5 145


Developing Expert Leadership for Future Schools


Experiences aimed to develop leaders for future schools ought to include oppor-tunities to appreciate and carry out the strategies for fostering more collaborativeschool cultures, as discussed in this chapter. Such strategies encourage and permitthe expertise of colleagues within the school to serve as powerful stimulants forteacher development.

Part 3

Developing Expert Leadership forFuture Schools


Chapter 10

The Socialization of School-Leaders

The term 'socialization' conveys quite different meanings to people in differentcultural settings. To some, the term is almost synonymous with brainwashingor, at thc least, 'swallowing' in an uncritical fashion the purposes and proceduresof one's organization or group of colleagues. To be socialized, from this perspect-ive, is to become simply a cog in the organizational wheel, a mindless bureaucratserving the policies of the organization, which are interpreted in a literal fashionand without a sense of personal empowerment.

In this chapter, the term socialization is intended to convey a quite differentmeaning. After Merton (1963), the term socialization encompasses those processesby which an individual selectively acquires the knowledge, skills and dispositionsneeded to perform effectively the role of school-leader. Such processes may rangefrom carefully planned, formal education programs, for example, through lessformal but still planned experiences (e.g., working with a mentor), to informal,usually unplanned on-the-job leadership experiences. One need not become a cogin the organizational wheel as a result of experiences like the. Indeed, it :spossible to imagine such processes fostering the individual's critical capacity ashe or she gets 'clowr' to the organization and better able to see blemishes onthe corporate body. Whether or not 'getting closer' magnifies thc corporatcblemishes and encourages a disposition toward change, or blurs one's sight andcreates blind acceptance, also depends on the nature of the socialization process.The ideal socialization process positions one at the point of sharpest focus: not soclose as to render the corporate nnage a fuzzy blur; nor so far away as to makethe detailed features of the image unrecognizable.

As we have argued in Part I of this book, leaders of future schools will needto possess a detailed, well focused picture of their school organizations. They willalso require a capacity to envision ever more productive forms that those schoolsmight take and the disposition to embark on changes, driven by a desire toapproximate those more productive forms. The overriding question of interest inthis chapter is: What kinds of socialization experiences contribute most to thedevelopment of such qualities in school-leaders (effectiveness on the high groundand expertise in the swamp)?

Together, Tables 10.1 and 1(1.2 summarize the framework for the chapterand the research on which it was based. This framework took as its point ofdeparture results of the small body of theory2 and empirical research' relevant to

1 6S.

Table 10 I. Socialization patterns experienced by aspiring principles

Dimensions ofSocialization Experiences Not Helpful

Socialization PatternsVery Helpful

1 Relationships(a) Superordinates

MI Peers

2 School-SystemPolicies, Proceduresand 'Control Mechanisms'

3 Formai Training

4 0 utcom esImage of Role

lc) Norms and Values

(Aspiring Principal) discouraged from entering role,exposed to models of ineffective practice; nocommunication networks developed for organizationalproblem-solving

interacts with peers primarily about social matters,has infrequent or no opportunities to assumeleadership roles with peers on school-wide and anddistrict issues influencing schools

establishes a reward system le g., promotion) inwhich merit is frequently subordinated in favor ofloyalty, personal relationships and other 'political'criteria

university-sponsored field-experience is shOrt-termand perfunctory

adoption of image of role as 'building manager' bythose not in role, lack of clear image of role by thoserot in the role

linimal development of skills required for entry to

decreased sharing of values and norms of behaviorwith super,ors

(Aspiring Principal) receives active sustainedencouragement and sponsorship to enter role;exposed to models of effective school administration,increase in the organization's communicationnetwcrks which foster more successful problem-solving- interacts regularly with peers about classroom,school-wide and district issues; has manyopportunities to assume leadership role with peers onschool-wide and district issues influencing schools,develops a network of peers sharing aspirations forthe principalship

establishes a reward system which acknowledgesmerit through promotion

focused on the understandings required to exerciseinstructional leadership, opportunities to reflect onrelationship between own school problems andknowledge provided in formal training programs,opportunities to experience the world of the principalfirst-hand over an extended time period; availability offrequent short programs focused on particular skillsand issues, opportunities provided by district to learn'how things are done' in the district

adoption of image of role as instructional andcurriculum leader

- significant development of skills required for entryto role- increased sharing of v,lues and norms of behaviorwith superiors

169 170

T.tble 10 2 Socolozaton patterns exper.enced by pract,c,ng pralopals

mes,ons ofSoc,allat,on E,cper e^r_es


S .tvy1 -ates

ipi SP:'

Socialization PatternsNot Helpful Very Helpful

(Practicing Principal) receives feedback whenschool problems become visible, exposed tomodels of ineffective practice in both school andschool-system leadership, communicationnetworks developed are designed to protect self-interest and school interests (damage control) in theface of system-wide decisions

occasionally interacts with peers about district:ssues, has occasional opportunities to assumeleadership role with peers on district issuesnfluencing schools, develops a network of peers toprovide advice, as needed

- has few requests for advice from subordinatesasp,ring tO the school administravon, remains distantftom the perspective of subordinates on classroomand school-wde matters

most contacts with students concern disc,piine andscHool-tout,ne issues

has nO strateg.es to foster self reflection

(Practicing Principal) - receives continuousfeedback, support, collaborative planning, andclear, realistic expectations, exposed to models ofeffective practice in both school and school-wideleadership; increase in the organization'scommunication networks which foster moresuccessful problem-solving

regularly interacts with peers about social,classroom, school-wide and district issues; hasmany opportunities to assume leadership role withpeers on district issues influencing schools;develops a network of peers to provide advice andcoat.hing

encouraged to develop a mentoring relationshipwith subordinates who show potential for schooladministration, actively encouraging theirdevelopment, has regular opportunities to betterappreciate the perspectives of subordinates onclassroom and school-wide matters; receivesregular feedback from experienced teachers abouthow they are doing, works with teachers onschool-improvement projects

maintains close contact with students

encounteredmaintains case records of significant problems

2 Schoot-Systern Pollutes andPocedures

3 i-orma, ra,n!ng

4 Outcomesimage 04 Ro P



1 3

explicitly focuses close to half the principals' timeon non-instructional matters in the school, severelyconstrains opportunities to address school-basedneeds, giving the major priority to system-wide needsand rewards reinforces a 'building manager' image ofthe role

no ongoing training available routinely in the schoolsystem and little effort made to find traintngelsewhere

- Increased commitment to tmage of role as 'buildingmanager'

refinement of a narrow range of school-management skills not directly related to schoolimprovement

decreased sharing of values and norms withsuperordinates, lack of procedural skills in day-to-daymanagement of school, Insufficiently developed skills

interpersonal relations, lack of knowledge abouthow school system really operates and whatprocedures to follow

- focuses a large proportion of the principals' timeon school improvement, encourages principals toaddress school-based needs in conjunction withdistrict-wide needs, consistently reinforces andrewards an 'instructional leadership' image of therole: encourages greater experimentation and risk-taking by principals, provides opportunities formentoring, apprenticeships, and the shadowing ofother principals

multiple opportunities available and taken, whichprovide access to increased skill, currenteducational knowledge, and attitudes ofcontinuous learning

- increased commitment to image of role asinstructional and curriculum leader

significant refinement of existing skills andaddition of new skills especially useful for schoolimprovement

increase in values and norms hared withsuperiors

1 4

Developing Expert Leadership for Future Schools

the socialization of formal school-leaders. Table 10.1 summarizes differences incocialization processes experienced prior to appointment as a vice-principal orprincipal. Table 10.2 is directed at socialization processes experienced after selec-tion as a formal school-leader.

Four dimensions of socialization are included in the framework. The firstdimension is the nature of the experiences provided by the relationships engagedin by those being socialized. These are experiences provided, for example, byrelationships with peers, superordinates and subordinates (for those already in therole) in the school district. Experiences resulting from contact with districtpolicies and practices is the second dimension. Especially relevant to this dimen-sion are policies and practices related to administrator selection and promotion,performance appraisal, and professional development. Formal education experi-ences and outcomes of socialization experiences are the final dimensions of thefra mework.

As Figures 10.1 and 10.2 indicate, the small amount of theory and empiricalevidence which is available suggests that there are, at present, different patternsof socialization experiences and that each pattern has different effects on thosebeing socialized. Such differenccs arc summarized in a bipolar way as being'not helpful', and 'very helpful' in developing qualities associated with effectiveschool-leadership. The 'very helpful' socialization pattei n contributes to theadoption and maintenance of an instructional leadership image of the role muchlike the program manager and systematic problem-solver described in Chapter 4;it fosters, as well, the development and refinement of knowledge and skillsrequired to carry out such curriculum and instructional leadership. Very helpfulsociahzation contributes to consensus about central organizational norms andvalues among administrators, at the school and district levels.

Taken together, Tables 10.1 and 10.2 also acknowledge that there are pre-dictable stages in the socialization process and that the perspectives and needs ofpeople may well differ depending on the stage in which they find themselves.Ronkowski and lannaccone's (1989) review of studies using a conceptualizationof socialization stages (initiation, transition and incorporation), drawn from thework of van Gennep (1960) and from Becker and Carper (1956), supports theusefulness of such a conceptualization. At the initiation stage people are con-cerned primarily about how others judge their adequacy. The standard forcomparison at the transition stage is some sense of required job performance. Atthc incorporation stage comparisons are with oneself (how much 'I' have changedfrom my previous self toward becoming an effective school-leader). Table 10.1 isintended to describe forms of socialization primarily relevant to those somewherein the initiation and transition stages of their socialization. Table 10.2 is relevantto those at the incorporation stage.

Given this framework, our general interest in this chapter about the mostuseful forms of socialization seems best served by seeking answers to four speci-fic questions: Do school-leaders experience significantly different patterns ofsocialization? If so, what are the effects of these different patterns? Which specificsocialization activities do school-leaders perceive to be most useful in their owndevelopment? Are different career paths associated with different socializationexperiences? Differences related to gender, stage of socialization, and district orregion also will be explored as part of several of these questions. It is worthnoting that the results of the research described below bear most directly on the


1 'jf-1t)

The Socialization of School-Leaders

development of effective leadership on the high ground. More tenuous, fromthese results, is the link between what we describe as helpful forms of socializa-tion and expertise in the swamp. That link needs considerabl) more scrutiny thanwe are able to give it in this chapter.

Answers to the specific questions addressed in the remainder of this chapterare based on the results of a series of four related studies. Results of the third ofthese studies are used as the primary evidence in answering the specific questionsof interest in this chapter. Study One (Leithwood, Steinbach and I3egley, inpress) used a thirty-two item survey-questionnaire to collect data from a sampleof forty-nine aspiring and thirty-eight practicing school-leaders drawn frommany different school systems in Ontario. Study Two (Leithwood and Stein-bach, in progress) was based on lengthy interviews with twenty-six practicingschool-leaders, probing, in some depth, questions left uncertain or unansweredby the first study: these school-leaders were also drawn from many differentschool systems, all in Ontario.

The third study, of primary interest in this chapter, was a survey, as inStudy One, with a forty-item instrument refined and extended to take intoaccount what had been learned in the two previous studies. In order to examinepossible regional differences, respondents were practicing school-leaders (princip-als and vice-principals) from one school system in the eastern Canadian provinceof Nova Scotia (113 respondents), one school system in the central Canadianprovince of Ontario (seventy-eight respondents) and a group of twenty-sevenrespondents from a number of school systems in British Columbia, the mostwesterly Canadian province.

The fourth study (Begley. Campbell-Evans and Brownridge, 1990) wasconducted in thc context of a school-leader preparation program jointly de-veloped and run by the Northwest Territories Departnient of Education and theOISE Centre for Leadership Development. Eighty-seven aspiring and practicingschool-leaders enrolled in this program, provided material for the survey, actingas participant observers, and supplied journal data to the study. Survey andinterview techniques were combined for the last study in this series with aspiringand practicing school-leaders in one large school system in south central Ontario(Leithwood et al., in progress). The discovery from previous studies of signi-ficant district effects on school-leaders' socialization experiences and a desire toredesign such experiences through some action research was the reason forconducting this study within a single school system. Subsequent results aredrawn from Study Three unless otherwise noted.

Question 1: Do School-Leaders Experience Different Patterns ofSocialization?

In order to inquire about differences in administrators' socialization experiences,each respondent was first classified depending on his or her opinion on wcializa-non activities, as having a pattern of socialization that was either 'not helpful','mixed', 'moderately helpful', 'high mixed', or 'very helpful'. This classificationwas based on responses to ten questions. Each of these questions concerned asingle socialization activity (e.g.. extent of exposure to effective role-models) andrequired respondents to identify, from a fixed set of alternatives, the nature of



Developing Expert Leadership for Future Schools

that activity as they had experienced it. These alternatives ranged from thoselikely to contribute least (rated as 1) to the respondents' administrative develop-ment to those likely to contribute most (rated as 3), as suggested by Tables 10.1and 10.2. Similar ratings on six or more of the ten items qualified a person to beassigned to one of three socialization patterns: six or more ratings of 1 led toassignment in the 'not helpful' pattern and a similar number of ratings of 2resulted in assignment to the 'moderately helpful' pattern. Assignment to the'very helpful' pattern required ratings of 3 on six or more of the ten items. The'mixed' patterns were reserved for respondents who did not rate at least six itemsin a similar manner. Those assigned the 'high mixed' category had five items ofthe ten rated as 3, with at most one item rated as I.

Table 10.3 identifies the ten items used for determining socialization pat-terns, reports the means and standard deviations for responses in Study Three toeach item by those assigned to each pattern, and indicates the number of peopleassigned to each pattern. The number assigned to each item corresponds to thequestion about this item as it was numbered in the survey instrument (the samepractice is followed for Tables 10.3 and 10.4).

As the last line of Table 10.3 indicates, only nine people (4 percent) appearedto have experienced the 'not helpful' socialization pattern. Relatively few people(eighteen or 8 percent) reported a 'very helpful' pattern, although when thesepeople are combined with those experiencing the 'high mixed' pattern (t wenty-four), almost a fifth of the respondents appear to have had relatively positiveexperiences. The bulk of respondents (167 or 77 percent) were classified as havingreceived a 'moderately helpful' (42 percent) or 'mixed' (35 percent) pattern. Theseresults are very similar to findings in Study One, with a somewhat smallerproportion of respondents reporting 'high mixed' or 'very helpful' socializationexperiences about 5 percent fewer.

Sources of difference in the socialization experiences of respondents areevident in all of the items in T,Ible 10.3. These results are very similar to thosefound in Study One; in that study, 'opportunity to interact with peers' (item 11)was the only experience in which there did not seem to be differences acrosssocialization patterns. Ratings for the same ten items were also examined forpossible differences related to gender and stage of socialization. Statisticallysignificant differences (p < (1.05 = the probability of these differences occurringby chance is less than 5 in 100.) were evident in the responses of women and mento two items. Men reported receiving encouragement to consider the role ofschool administrator earlier in their careers than did women (item 7), a trend alsoevident in Study One. Unlike Study One, women reported more frequentopportunities to assume leadership roles on school or board-wide committees(item 22). Study Three did not find (as was the case in Study One) a tendency forwomen to report greater participation in a network of peers also aspiring LO

school-leadership positions (item 19). Two statistically significant differenceswere also evident when the responses of those in the earlier stage of theirsocialization (vice-principals) were compared with those in the later stage(principals). Principals reported earlier encouragement to consider a school-administration role (item 7). Vice-principals reported more frequent exposure toeffective role-models (item 15), a difference favoring women over men in StudyOne. Study One reported statistically significant differences favoring principalson items related to the intent of board hiring and promotion practices, evaluation





Table i 0 3 Patterns of school-leaders socatzation experiences

Socialization Activities

t6! Extent of encouragef,leof topursue adminsirator loie

I!) Stage .n career encourapernentreceNed

f x:/-ff. Of ,Avoqu,-uty to ntea,:zpe

'5, 1 PTI if to effect.vee rrode.s

r (1i pa!! c neer"W.Vvor ()' asVra'1:s

g :or .; O oppoffur:ty 10 assur-neeadersh.p roies

,/i), f ffer:r " ,-1,0 of ur"ol Dr,

, ,-fer't aln, s'rafr evadlorP',1 H es

1 (1, Pirlt e s reurr4f 0"

Socialization PatternsMean ratings for each achvity per group (scale. 1-3)

Not Moderately High Very

He)pful Mixed Helpful Mixed Helpful

Mean S D Mean S D Mean S.D. Mean S D Mean S.D

Nuf ;be, of respondents in eachpattern ipercemage of

1 56 0 88 2 24 0 77 2 18 0 60 2 67 0 48 2.71 0 47

2 00 0 87 2 24 0 77 2 28 0 61 2 75 0 44 2 82 0 39

1 89 0 60 2 37 0 54 2 35 0 48 2 75 0 44 2 94 0 24

1 6; 1 00 2 63 0 57 2 02 0 53 2 29 0 46 2 71 0 47

1 33 0 60 1 88 0 781 22 0 44 1 64 0 59 I 96 0 55

1 67 1 00 2 24 0 62 1 97 0 56 1 44 0 51 2 88 0 33

1 f-i) 0 55


1 41), 0 69


1 22 0 6- 2 72 0 67 2 18 0 73


1 60 0 66 2 11 0 38-.iv

1 11 0 33 1 91 0 33 E.'2 35 0 61


2 1:i 0 69


1 25 0 It 1 52 0 63 1 83 0 51 2 00 0 87 'Z...


; 79 0 86;.--

1 44 0 73 2 14 0 70 2 83 0 48 2 88 0 33 PS.


9 14%1 76 135%! 91142%1 24(11%) 18 (8%) -

Notes Data estimate the degree of helpfulness of socialization activities using results of pievious research about effective socialization experiencesr7 9

,. n Devat,on


Table 10.4: Relationship between importance of socialization activities and socialization patternsSocialization Activities

Socialization Patterns Mean ResponseMean ratings for each activity per aroup (scale 1-4) for Each Activity

Not Moderately High VeryHelpful Mixed Helpful Mixed HelpfulIN = 91 IN = 76) IN = 91) (N = 24) IN = 181Mean S D Mean S D Mean S D Mean S D Mean S D191 Encouraged to pursue 3 40 1 34 2 93 0 99 3 01 0 85 3 65 0 57 3 17 0 99 3 23administrator role

1121 Opportunities to rnteract 2 67 1 41 2 81 1 03 2 97 0 97 3 52 0 5c 3 83 0 38 3 16with peers

(14) Having a mentor 3 50 0 71 3 39 0 86 3 17 0 78 3 50 0 71 3 86 0 36 3 48116l Exposure to effective 3 00 1 31 3 04 0 96 3 U9 0 75 3 25 0 79 3 78 0 43 3 23role niodels

1171 Preparation for teaching 1 88 1 13 1 80 1 00 1 69 0 92 1 86 1 04 1 75 0 93 1 80(181 Teaching experience 2 43 1 13 3 09 0 95 3 00 0 90 3 14 0 83 3 18 1 02 2 971211 Participation in peer 2 33 0 58 2 92 1 10 2 53 1 08 2 94 1 03 3 18 0 75 2 76network of aspirants

1231 Opportunities to assume 3 GO 1 41 3 03 0 86 2 86 0 88 3 30 0 77 3 94 0 24 3 23leadership roles

127) Peiception of hiring and 1 29 0 76 1 89 1 05 2 25 1 13 2 55 1 06 2 63 1 31 2 12Dicrnotn practices

1291 Perception of administrator 1 14 0 38 1 30 0 61 1 67 0 94 1 91 1 00 1 88 1 05 1 58evaluation practices

1311 Administrator-preparation 2 88 0 76 2 69 0 77 2 62 0 70 2 55 0 53 2 46 0 52 2 64programs

1341 Overall perception of 3 33 I 00 3 42 0 76 3 46 0 62 3 65 0 65 3 75 0 45 3 52readiness

Mean Response for 2 50 0 80Each Patterm

2 55 0 47 2 62 0 48 2 92 0 25 3 06 0 40

Note Data estimate the value of socialization activities using respondents perceptions'N Number of respondents

The Socialization of School-Leaders

practices, and preparation programs (items 25, 28, 30). Such trends were alsoevident in Study-Three dat3 but the differences were not statistically significant.

Although not evident from Table 10.3, about twice as many principals asvice-principals (25 percent vs. 13 percent) reported the two most helpful patternsof socialization. This difference is largely attributable to the Ontario samplewhere 55 percent of principals compared with 36 percent of vice-principalsclassify their socialization experiences as either 'high mixed' or 'very helpful'.Study One also found similar differences; in that study, the comparison wasbetween aspiring and incumbent school-leaders, with incumbents generally re-porting more helpful patterns of socialization. Also relevant, but not evidentfrom Table 10.3, were differences in the patterns of socialization experienced bythose in different districts and/or regions of the country. A much larger propor-tion of respondents from the Ontario school system reported either 'high mixed'or 'very helpful' patterns of socialization than did respondents from other re-gions: 48 percent of Ontario respondents as compared with 12 percent in BritishColumbia and only 2 percent in Nova Scotia. The two least helpful patterns ofsocialization were reported by 23 percent of the Ontario sample, 37 percent of theBritish Columbia sample and 50 percent of the Nova Scotia sample.

Do school-leaders experiences different patterns of socialization? The answerfrom Studies One and Three, in sum, is: definitely yes. These studies len,-.1support to claims that:

Most aspiring and practicing school-leadcrs experience a 'moderatelyhelpful' pattern of socialization (in their own opinion); few experience auniformly negative socialization pattern whereas 19 percent experience aquite helpful pattern (High Mixed + Very Helpful)District effects on socialization experiences are very strong; differences inactivities at the district level may be capable of determining whethervirtually no school-leaders or as many as 48 percent count their socializa-tion experiences as relatively helpful in their own development (HighMixed + Vcry Helpful)Women and men experience very similar socialization patterns althoughmen appear to receive earlier encouragement to consider the role, where-as women perceive more frequent leadership opportunities available tothemSchool-leaders in the later stage of their socialization generally perceivehaving experienced a more helpful pattern of socialization than those inthe earlier stage

Question 2: What are the Effects of Different Patterns ofSocialization?

Among the most important reasons for our interest in socialization patterns is theexpectation that differences in such patterns account for a substantial amount ofthe variation observed among school-leaders in their orientation to the role andeffectiveness in it. Increasingly helpful patterns, as described in this chapter,should contribute to an orientation to the role generally consistent with what wasdescribed in Chapters 4 and 5 as effectiveness and expertise. To inquire about this



Developing Expert Leadership for Future Schools

relationship, it was assumed that indicators of a school-leader's orientation to therole could be found in responses to four open-ended questions. These questionsconcerned `overall image of the role', `most important tasks' associated with therole, and tasks in which respondents were most and least confident in their abilityto carry them out.

In order to interpret responses to the question about overall images of therole, the four-fold conception of dominant orientations of school-leaders toproblem-solving on the high ground, described in Chapter 5, was used im:ially.Eventually, we combined the two orientations considered most effective on thehigh ground and those considered least effective.

Overall, 44 percent of the total sample of respondents (as well as theprincipal and vice-principal sub-samples) identified images of the role, consistentwith the most effective patterns of leadership. As with Study One, results ofStudy Three suggested that as socialization patterns became increasingly helpful,larger percentages of respondents expressed images consistent with the mosteffective orientations to problem-solving. 11 percent, 39 percent and 45 percentof respondents experiencing the three least helpful patterns held such images, incomparison with 58 percent and 56 percent of those who had experienced thetwo most helpful patterns ('high mixed' and 'very heipful'). Among the 19percent (forty-two) of respondents who reported the two most helpful socializa-tion patterns, 69 percent (twenty-nine) were principals and 31 percent (thirteen)were vice-principals. About half of these principals but more than three-quartersof these vice-principals expressed images consistent with the most effectiveorientations to problem-solving.

A content analysis of responses to the open-ended questions concerningmost important tasks and tasks which one had most and least confidence resultedin the identification of nine categories of tasks. Five of these tasks were directlyassociated with the emphasis one would associate with effective problem-solvingon the high ground: instructional leadership; having a student focus; effectiveadministrative problem-solving; providing leadership (e.g., role modeling, moti-vating, facilitating); and continuing to acquire new knowledge. Three of theremaining four tasks were primarily interpersonal in nature: communication;developing a positive climate; and personnel relations (e.g., staff development,being supportive). A final task identified was the management of routines (e.g.,budget, discipline). (To be clear, and as we explained in Chapter 5, ail of thesetasks are required of school-leaders. But the preoccupation of less effectiveschool-leaders tends to be with tasks like the last four described above. Highlyeffective leaders, in contrast, carry out such tasks but consider tasks like the firstfive mentioned above to be priorities for their attention.)

Based on previous research, one might expect those experiencing increasing-ly helpful socialization patterns and espousing images of the role consistent witheffective orientations to problem-solving, to award more importance to thosccasks directly associated with such problem-solving. Those experiencing the twomost helpful patterns differed only marginally from those experiencing the threeleast helpful patterns, however. Each set of respondents identified about equalnumbers of tasks associated with effective problem-solving and not so associated,a finding similar to Study One. Further, there were no regional differencesevident in these results. In a similar fashion, respondents, experiencing the twomost helpful patterns combined. had less confidence in their ability to carry out

1 58

18 3

The Socialization of School-Leaders

tasks preoccupying less effective problem solvers as compared with tasks ofpriority to highly effective problem solvers; 72 percent vs. 28 percent in the caseof least confidence and 60 percent vs. 40 percent in the case of most confidence.Respondents who experienced the three least helpful socializations patterns re-ported very similar tendencies: identical in the case of most confident tasks and 63percent vs. 37 percent in the case of least confident tasks. These results are alsosimilar to the results of Study One. Differences based on region or district werevery small.

Among all tasks, respondents expressed least confidence in their ability toperform what we have been calling 'managerial' tasks. Included in this categorywere such matters as office procedures, discipline of students, report writing,processing paper work, and interpreting legal acts and regulations and the like.About 34 percent of all statements, made about least confident tasks by the totalof 157 who responded to this question, as well as by Ontario and Nova Scotiarespondents, concerned this category. A much smaller proportion of BritishColumbia principals (10 percent) demonstrated this pattern.

Acknowledging the correlational and, therefore, only suggestive nature ofthe evidence, what are the effects, in sum, of different patterns of socialization?Results of our research (both Studies One and Two) suggest that:

There is a predictable relationship between school-leaders' images oftheir role and the patterns of socialization which they experience.Increasingly helpful patterns are associated with a tendency to adoptimages of the role, consistent with effective forms of school-leaderproblem-solvingIndependent of the socialization pattern experienced and image of therole adopted, school-leaders award about equal importance to inter-personal and managerial tasks, as to tasks more central to effectiveproblem-solving on the high ground. These interpersonal and managerialcategories of tasks also contain a higher proportion of specific tasks inwhich school-leaders are least confident. Managerial tasks are a source ofgreatest uncertainty for all school-leaders. Thc region or school districtin which school-leaders work does not seem to affect these results.

Question 3: Which Socialization Activities Do School-LeadersPerceive to be Most Helpful?

Data reported in Table 10.3 estimate the degree of helpfulness of selectedsocialization activities using, as the standard, results of previous research abouteffective socialization experiences. Data summarized in Table 10.4 help estimatethe value of most of these same socialization activities, using respondents' percep-tions rather than previous research results as the standard. These perceptions areuseful in helping to confirm or disconfirm results of previous research. So, forexample, if respondents who indicated receiving sustained early encouragementth become administrators also indicated that it was not important in their deci-sion to pursue the role, that would be considered a challenge to the validity ofprevious research tindmgs concerning effective socialization activities.



Developing Expert Leadership for Future Schools

Table 10.4 reports the school-leaders' perceived importance to their adminis-trative development of the socialization activities appearing in Table 10.3 (plusthree additional factors: teaching experience, preparation for teaching and havinga mentor). Respondents' perceptions of their overall readiness for an administra-tive role is also reported (respondents rated importance on a four-point scale, 4being very important, 1 being not important).

In general, support for previous research results concerning the value of asocialization activity would be evident whcre ratings increased, as socializationpatterns became more helpful. Such an interpretation of the data seems war-ranted, based on the mean ratings of all activities assessed by respondents (`MeanResponse', Table 10.4). While respondents in the 'not helpful' pattern providedthe lowest mean rating (2.50), there are virtually no differences in mean ratingsfor respondents in the 'mixed' (2.62), and 'moderately helpful' (2.62) patterns.Those respondents in the 'high mixed' and 'very helpful patterns', however,awarded greater value (2.92 and 3.06 respectively) to their overall socializationexperiences, as would be predicted by their generally more helpful socializationpattern; they also perceived themselves to be readier for the challenges of schooladministration (item 34). These results replicate findings from Study One.

Support for previous research results is also provided in relation to fourspecific socialization activities: exposure to effective role models, having a men-tor, opportunities to interact with peers, and opportunities to assume leadershiproles. Thcy were seen, on average, as being very important aids to leadership de-velopment. Across all socialization patterns, opportunities to assume leadershiproles were rated as most helpful, a finding similar to evidence reported in StudyOnc. Among the twelve socialization activities reported in Table 10.4, ratings forsix fell below 3.0: preparation for teaching (Mean Response = 1.80), teachingexperience (M.R. = 2.97), participation in peer networks (M.R. = 2.76), percep-tion of administrator evaluation practices (M.R. = 1.58), administrator pre-paration programs (M.R. = 2.64) and perception of hiring and promotionpractices (M.R. = 2.12).

Differences between women and men in the perceived importance of thetwelve socialization activities were also examined. Having a mentor (item 14)was rated highest by both groups. Other activities, rated highly by both, in-cluded encouragement to pursue the role (item 9), exposure to effective rolemodels (item 16), and opportunities to assume leadership roles (item 23). Womenrated seven of the twelve activities higher than did men, but none of thedifferences were statistically significant. In spite of this, women rated theiroverall readiness for the position lower than did men. This difference approachedstatistical significance (p = (1.078) and reinforces the statistically significant resultreported in Study One.

Those in the earlier stage of their socialization (vice-principals) attributcdmuch greater importance to teaching experience (item 18) although this differ-ence did not quite reach statistical significance. Mean ratings, given the twelvesocialization activities, were marginally higher overall for those in the earlier, ascompared with the later, stage of their socialization, as was also the case in StudyOne.

Data reported in Table 10.4 give sonie indication of the perceived value ofonly a small number (twelve) of socialization activities. However, the surveycontained four open-ended questions bearing on this issue as well. These clues-



The Socialization of School-Leaders

dons asked respondents to list other people, experiences or factors contributingto the aspiration to become a school administrator, identify most and least helpfulexperiences, and describe experiences which they would recommend to others.Responses to these four questions were summarized around categories whichemerged naturally from what was written on the surveys.

As in the earlier study, specific, on-the-job leadership activities were cited asmost helpful more frequently than any other category of activity (29 percent ofall responses). Having experience in as many professional settings as possible(breadth of experience) was viewed as the second most helpful kind of activity(25 percent). Taken together, these two categories account for 54 percent of allresponses and suggest that these respondents feel that experience is the bestteacher. However, the experiences must be seen as worthwhile since they canalso be counted as relatively useless when they are meaningless and trivial (e.g.,counting books). For example, breadth of experience and on-the-job leadershipexperience were considered to be the second and third least helpful kinds ofactivities (10 percent and 7 percent respectively). Respondents felt that formaltraining was the least helpful preparatory experience they had had, with 70percent of all responses falling in that category. Again, though, depending on thespecific form it takes, formal training can be seen as very helpful; it was the thirdmost helpful experience cited with 16 percent of responses.

In the same vein, relationships with superordinates was the fourth mostfrequently mentioned category on both lists accounting for 15 percent of re-sponses to the most helpful experience question and 7 percent of responses to theleast helpful question. The other most helpful experiences were personality fac-tors such as be a self-starter, be a risk-taker, have a sense of humor (6 percent);relationships with peers (6 percent); and personal history or family relationships(2 percent). District policies were never mentioned as being helpful to becominga school-leader but it accounted for 3 percent of all responses to the least helpfulquestion. Lengthy or poor interviews and lack of specific policies or initiationsare examples of the types of activities cited here.

The responses to the open-ended questions showed consistency across dis-tricts with one exception: in answer to the question regarding most helpfulexperiences, British Columbia was half as likely to report relationships withpeers, as the other two districts (British Columbia, 4 percent of responses; NovaScotia, 8 percent; Ontario, 9 percent).

Both Studies One and Three centered their inquiry about sources ofsocialization on what might be termed external sources sources outsidc theindividual. A useful perspective on the importance of these sources is providedby Study Four. In addition to a concern for external sources, this study alsoexamined the role played by 'factors' internal to the individual, particularly assuch factors influenced decisions to pursue a school-leadership career and thetraining required. Results of this study confirm an important role for the externalsources identified in Studies One and Three. However, internal sources, such asthe need for challenge, pursuit of knowledge and skill, interest in a new role, andthe nccd for additional responsibility, were even more frequently identified. AsBegley et al. (1990) also point out, these factors provided practicing school-leaders in Duke's (1987) study with thcir greatest sources of satisfaction and, atthe same time, were the main causes of thcm lea% ing the job (e.g., excessivechallenge). Study Four suggests that internal sourc...s may be especially influential



Developing Expert Leadership for Future Schools

in the decision to pursue a school-leader career and the attendant training:external sources of the sort evident in Studies One and Three may be of particu-lar importance after those decisions are made.

Based on the data described in this sectiol. of the chapter. the followingclaims about most helpful socialization activities seem warranted:

Sources most influential in giving rise to the decision to become aschool-leader and prepare for the role appear to include such internal orpersonal qualities as the need for challenge, a thirst for knowledge andthe likeThose forms of socialization valued most after the decision to pursue aschool-leadership position appear to be embedded in the context ofschool life. On-the-job leadership experiences and having broadly-basedschool experiences (e.g., holding a variety of teaching positions in va-rious grades) are seen as being very helpful activities; however, depend-ing on the specific form, they can also be of little useFormal preparation programs for aspiring and practicing administratorsappear to vary widely in their perceived value. They are capable of beingvery helpful or extremely unhelpful, presumably depending on theirquality. Such variation rmy also be the case, in a less pronounced way,for relationships with superordinates

Question 4: Are Different Career Paths Associated with Differences inSocialization Experiences, Gender, or Socialization Stage?

One question on the survey asked respondents to indicate the roles which theyhad occupied over the course of their careers to the principalship and to identifywith numbers the sequence in which they had occupied such roles. Optionsincluded: (a) job outside education; (b) elementary teacher; (c) middle-schoolteacher; (d) secondary-school teacher; (e) consultant, coordinator, or the like; (f)department head; (g) elementary-school teacher; (h) vice-principal; and (i) other(to specify).

Responses were classified in relation to four alternative career paths. Path 1included those who had begun their careers with a job outside education and hadthen occupied a series of in-school roles through to their current job. Path 2included those who had arrived at their current job through a series of exclusivelyin-school roles. Path 3 included people who had experience in a district-level role(e.g., consultant), irrespective of the nature and sequence of other roles whichthey had occupied. Path 4 involved taking a job outside of education in themiddle of one's educational career (e.g., leave an education position but returnlater). These four paths were chosen as the basis for analyzing responses becausethey seem likely to help develop different skills or to contain experiences leadingto potentially different perspectives on the principalship. For example, a joboutside education may provide one with an opportunity o appreciate the unique-ness of the school's responsibility to society; a coordinator's job may provideopportunities to understand the stimuli giving rise to district initiatives and toacquire significant curriculum-management skills.

Of interest in analyzing the data were dominant career paths, as well as the

162 187

The Socialization of School-Leaders

relationship between career path and socialization pattern, gender and stage ofsocialization. Results indicated that Path 2 (in-school to administration) was thedominant pattern including 45 percent of respondents. Paths 1 and 3 were eachfollowed by about 25 percent of respondents. Only 5 percent of the respondentsfollowed Path 4. Paths 2 and 4 respectively were the most and least dominantpatterns found in Study One.

As in Study One, also, there was a distinct tendency for those experiencingmore helpful socialization patterns to have career paths including some out-of-school role, about three-quarters (74 percent) of those experiencing 'very helpful'or 'high mixed' patterns, as compared with less than half of the rest (46 percent).Men and women did not demonstrate large differences in their career pathsalthough a larger proportion of men than women followed Path 2 (48 percent vs.36 percent); the reverse was the case with Path 3 (23 percent vs. 36 percent).These tendencies were in evidence more dramatically in Study One. Only mar-ginal differences in career path were apparent between vice-principals and prin-cipals, replicating Study One results for those at earlier and later socializationstages.

One difference in career path appeared to be related to the district or region.The Ontario sample of respondents demonstrated a much greater tendency tofollow Path 3, involving some district-level experience (54 percent in Ontario vs.8 percent in Nova Scotia and 19 percenE in British Columbia). Such a pattern isnot surprising, given the encouragement in the district to consider such experi-ences as extremely valuable prcparation for school-leadership. This career pathemphasis in Ontario was mostly offset by a disproportionately small number inOntario (13 percent vs. 34 percent in Nova Scotia and 33 percent in BritishColumbia) following Path 1.

These results suggest, in sum, that:

Once aspiring administrators and principals begin their careers in educa-tion they rarely experiment with careers in other fieldsThe most frequently chosen career path to the principalship includes

in-school roles onlyA significant minority of aspiring administrators and principals have hadcareer experiences outside schools, either in district roles (more often forwomen) or roles outside education (more often for men). These careerexperiences are related to socialization experiences more helpful in pre-paring administrators for instructional leadershipDistricts are able to use career paths as preparation for school leadershipthrough conscious and visible attention to leadership-selection criteria.

Conclusions and hnplications

By way of conclusion, we explore some of the implications of two issues centralto our study: variations in the helpfulness of different socialization patterns andthe socialization experiences of women.

Socialization patterns experienced by respondents in the study were at leastmoderately helpful in contributing to instructional leadership. While this is good

news, it is also clear that there is much room for improving the quality of

1 S S163

Developing Expert Leadership for Future Schools

socialization experiences. Since such experiences seem likely to account for muchof the variation in the quality of leadership available to schools, they are a potentvehicle for school improvement. Furthermore, it is possible, with the greatlyincreased demand for new school administrators over the next decade (due toretirements), that socialization experiences may actually deteriorate. This is likelyif previous amounts of attention and effort given to socialization are spread acrossmany more people. One possible explanation for our finding that administratorsin the later stages of socialization experienced more helpful patterns than those inearlier stages is that such dilution may be occurring already.

Several suggestions for improving the socialization experiences of aspiringand practicing principals emerge from this study. One suggestion concerns for-mal training programs. Greenfield (1985) has suggested that such programs arethe primary vehicle for developing the technical knowledge and skill whichadministrators require. Yet their quality and impact is extraordinarily uneven(see, for example, Leithwood, Stanley and Montgomery, 1984, and Leithwoodand Avery, 1987). Devoting more time to programs which deliver substantiallymeaningful content, in a form consistent with good principles of adult education,is one promising suggestion for improving socialization experiences. The charac-teristics of such programs are examined in the subsequent two chapters. It is alsoworth noting the results of our research which indicated a lack of confidence inperforming managerial tasks; this might be due to our zeal to focus school-leaders on 'educational' aspects of their role, which would cause their managerialpreparation to be badly neglected.

A recent study by 1-'..pke (1989) supports thc claim, made in this chapter,that on-the-job leadership activities are viewed as among the most helpful of allsocialization activities. Our second suggestion is designed to make this activityeven more productive. Papke (1989), among others, proposes to do this byensuring that principals and vice-principals negot' IL . job responsibilities so thatvice-principals have experience with a comprehensive array of principals' respon-sibilities, not simply the mundane, routine, maintenance tasks. We recommendtaking this a step further by formally including in the criteria, used to evaluateprincipals, responsibility for the leadership development of their vice-principals.

On-the-job leadership experiences could be increased in their potency inseveral additional ways. For sonic school districts, routinely favoring the selec-tion of administrative applicants with experience in curriculum-consulting roles isthe most promising. As Ross (1989) has also observed, such roles offer uniqueopportunities to acquire many of the attributes associated with instructional leader-ship. These include, for example. the development of curriculum-managementskills, the reliance on authority grounded in expertise rather than position, refine-ment of communication and other interpersonal skills, development of groupproblem-solving processes, and acqeisition of a more comprehensive perspectiveregarding district-school relationships.

Finally, on-the-job leadership experiences could become more potent, bysystematically selecting those who have expressed an interest in school adminis-tration, to chair district-level committees working on any problem which re-quires the exercise of leadership. In addition, it would he worthwhile to providethis opportunity for those who have not yet expressed an interest but who seemto be likely candidates for administrative positions. This tactic would not onlyencourage those who haven't yct considered such a position, but would also



The Socialization of School-Leaders

build confidence and enable them to make judgments about their suitability for,and interest in, administration.

The five studies alluded to in this chapter included at least equal numbers ofwomen and men. It appears that women are now preparing to become schooladministrators in record numbers although the Study-Three sample of existingadministrators was weighted more than three-to-one in favor of men. This is adramatic change over one decade, judged by the available evidence (e.g., Haven,Adkinson and Bagley, 1980). Results of our studies suggest that this may be due,in part, to the greater encouragement women are receiving to consider adminis-tration (although still not as early in their careers as men). Undoubtedly, thebroader social recognition of discriminatory practices and generally enhancedfemale career expectations account for much of this change, as well. Notwith-standing such rapid progress, these studies point to an anomaly worth noting.Whereas women generally considered their socialization experiences, as a whole,to be at least as helpful, if not more helpful, than did men, they expressed muchless confidence than men in their readiness to assume an administrative role. Onepossible explanation for this apparent contradiction may be found in gender-based differences in confidence about one's abilities in relation to tasks and rolestraditionally associated closely with only one gender. Linn and Hyde (1989), forexample, have demonstrated this phenomena in relation to mathematics andscience abilities, abilities traditionally associated more closely with males. Evenwhcn gender groups perform equally, evidence suggests that males are muchmore confident about such abilities. Because formal school-leadership roles havetraditionally been dominated by men, results of our studies may be indicating alack of self-confidence on the part of women, unrelated to actual administrativeability.

This explanation, however, has little practical import and is premised onassumptions consistent with what Shakeshaft. (1989) describes as the third of sixstages of research on women in educational administration: women as disadvan-taged or subordinate. We find more helpful a perspective on these data, consis-tent with Shakeshaft's (1989) fourth stage: women studied on their own termsFrom this perspective, it becomes, especially important to inquire further into thereasons for women's 'only modest' levels of self-confidence and to develop formsof support that address these reasons, to be provided in the transition stages oftheir socialization. Following this line of reasoning, it seems worthwhile toinvestigate also the role that personality characteristics in general (for both menand women) play in socialization.


1 We wish to acknowledge Cisco Magagula's help with statistical analysis.2 Theoretical sources used to develop the framework included: Greenfield (1985), Peter-

son (1986). Silver (1986), and Kline (1988).3 Empirical sources used to develop the framework included. Cooper (1989), Daresh

(1986), Marshall (1984). Crowson and Morris (1985).


Chapter 11

Characteristics of Formal Programsfor Developing Expert SchoolLeadership on the High Ground

In Chapter 10 we outlined the array of experiences which, at least potentially,shape the readiness of individuals for the practice of school leadership. Wereported a wide range of opinion with respect to the perceived relevance offormal preparation programs. For some school-leaders, experiences with suchprograms proved very helpful; for many others the experience was perceived as awaste of time. While these results may come as no surprise to many familiar withsuch programs, they ought to be viewed as alarming for at least two reasonsAlthough seriously underfunded, for the most part, formal preparation programsare sufficiently institutionalized in many provinces, states, and countries that theyare likely to remain a part of school-leaders' experience during the 1990sMoreover, in light of the hectic and unreflective context characteristic of theschool-leaders' work environments, such programs continue to offer a more orless unprecedented oasis of opportunity, free from the 'press for action' for thedevelopment of key skills, knowledge, and dispositions.

Given the probable continuation of formal programs and their significantpotential for leadership development, program quality should be a primaryconcern. To address this concern, as it bears on preparation for the high ground,we pursue answers to three questions in this chapter: Why have so many formalpreparation programs failed to live up to the expectations of their developers andclients? What accounts for the success of one set of formal preparation programswhich have been widely implemented and systematically evaluated? What guide-lines can be offered to those wishing to develop and implement formal, high-ground preparation programs for school-leaders?

Why Many Formal Preparation Programs Fail

Reasons for the failure of many formal preparation programs, most of which aimto develop high-ground expertise, have been described in some detail (e.gFlallinger and Murphy, 1991, Blum and Butler, 1989, Griffiths, Stout and For-syth, 1988, l'itncr, 1987, Leithwood and Avery, 1987). According to Murphy(1990), the weakness of many formal programs can be classified as attributable to


Formal Programs Jor Developing Expert School Leadership on the High Ground

either program content or program delivery For present purposes, we touch on

them briefly to set the stage for describing a program which addresses many of

these weaknesses.

Proqram-Content Weaknesses

Three weaknesses attributable to course content arc especially noteworthy. First,

such content often makes questionable contributions to school-leaders' school-

improvement abilities by virtue of thc outcomes it helps to achieve:

Encompassed by this criticism are programs in which such outcomes are

not convincingly linked to school improvement; depend primarily on

the expressed needs of participants; are entirely 'issues dependent', not

addressing the principals' role in the issue; and/or do not recognize the

scope of the principals' job as a whole. (Leithwood, Stanley and Mont-

gomery, 1984, p. 51)

A second content-based weakness is characteristic of courses organized along

a thematic or issues approach, or those which allow candidates relative freedom

in selecting their learning experiences. Such programs are less likely to be found-

ed on a particular image of the role which integrates and brings professional

coherence to the various bodies of knowledge, skills, and attitudes required.

Many university-based programs, for example, have been critiLized as an eclectic

selection of courses that add up to neither a clear image of the role nor a reliable

set of relevant skills. This is an important consideration for course candidates

who have little or no leadership experience to use as a basis for identifying and

selecting appropriate learning c;;periences. In contrast, basing a preparation pro-

gram on a comprehensive and intograted image of school leadership allows candi-

dates to move beyond just the mastery of discrete skills. For example, there is

little doubt that the development of time-management skills is important to

school-leaders, but aspiring leaders should be able to answer the question: 'Time

management, to what end?'A third weakness in the content of many programs is their failure to come to

grips adequately with the full scope of the school-leader's role. Given the con-

straints of time and place, and the complexity of the role, it is unrealistic to

expect any preparation program to fully prepare the aspirant for the position.

Presenting and maintaining a balanced program is difficult at the best of' times.

Some types of activities have traditionally tended to be over-represented in

preparation courses while others have been ignored or given short shrift. For

example, knowledge and skills related to building management functions, in-

fbrmation about legislated acts and regulations, and timetabling procedures are

usually well represented. This reflects thc 'rear-view' perspective of many course

designs which are based on traditional practices and expectations rather than

current organizational needs. Moreover, such traditional content is easier to teach

than sonie of the more pro-active, open-ended, higher-order skills such as,

implementing an entry plan when assigned to a new school, or developing

positive school culture.


1 ci

Developing Expert Leader<hip for Future Schools

Program-Delivery Weaknesses

Three shortcomings are especially evident in the delivery of many programs. Agreat many programs fail to consider the developmental aspects of the school-leader's role and particularly the varying stages of readiness manifested by in-dividual candidates. Program candidates vary considerably in prior experienceand most preparation programs are not sensitive to such variation. Begley,Campbell-Evans and Brownridge (1990), in their research on factors which in-fluence the socialization of aspiring principals in the Northwest Territories,have found that individuals enroll in preparation programs for a variety ofreasons, with a variety of prior experiences and qualifications, as well as variedexpectations for the program. More significantly, course candidates, who weresurveyed and interviewed in this study, apparently began their preparation pro-grams with vague or varied images of the principal's role.

A second consideration is that program participants, many without previousadministrative or leadership experience, cannot be expected to become full-blownschool-leaders simply by imitating the actions of course instructors who areexemplary practitioners or by learning sets of procedures passed on by instruc-tional staff, presenters or academics. The incorporation of a school-based practi-cum component as part of the preparation experience begins to address thisshortcoining. However, to be effective, skill-application experiences must beproperly supervised or coached and be of sufficient duration and substance toassure significant learning. This is often not the case.

A third prograin-delivery weakness arises out of a tendency, given theconstraints of time and place, for programs to be focused on what principals do,or at best, on generalized procedures for carrying out their responsibilities. Littleattention is typically devoted to encouraging reflection by candidates on whysuch actions are appropriate, and on the intent of such actions or the variations inapproach necessitated by situational factors. Several authorities on principal de-velopment, for example Barnett and Brill (1990), have become increasinglyinterested in building such reflection into administrative training programs.

One further characteristic of most formal programs, which perhaps trans-cends the categories of content and delivery, is what Gaines-Robinson andRobinson (1989) describe as the 'training for activity' trap. This trap arises whenprogram designers and implementors are held responsible, as is typical, for the'activity' of the program, but not for its results. Accountability in the trainingsector (whether education or industry) is restricted to such criteria as the numberof programs offered, the number of participants enrolled, and the relative cost ofprograms. Program designers become preoccupied with the design and deliveryof programs, leaving little or no incentive to do needs' assessments beforehand orresearch on program outcomes. Similarly, there is a frequent and equal absenceof identified management responsibility for the results of training programs.Ultimately, no one person or group has accepted accountability for ensuring thatparticular knowledge, skills or attitudes viewed as desirable will be applied by thccourse participants when they carry out their professional roles.

As a consequence of this lack of accountability, the degree of skill transferfrom the program context to the classroom and school is usually unknown orabsent. The primary concern of thc course implementor is providing a high-quality learning environment and producing high candidate satisfaction with the

168 191 A

Formal Programs for Lieveloping Expert School Leadership on the High Ground

program Program activities frequently lack a clear link or alignment with eitherwhat school administrators do on the job or the particular professional needs ofindividuals. Traditional university-based courses tied to graduate degree pro-grams add insult to injury, by not necessarily being sensitive to the canons of gdodpedagogy or candidate-satisfaction levels. As Gaines-Robinson and Robinson(1989) point out, course activities are more likely to reflect a stereotyped require-ment of a course, which has developed a life of its own through repetition, ratherthan any identified need expressed by candidates or perceived by the programsponsors. Furthermore, course candidates, who may or may not have need for aparticular skill activity, must typically participate in all activities because ofexpectations for a uniform-preparation course experience.

Clearly, the design and delivery of school administrator pre-service pro-grams is a complex business fraught with shortcomings and challenges. Subse-quent sections of this chapter give consideration to the characteristics of formalprograms most helpful to the development of leadership on the high ground.

Characteristics of Successful Programs: An Illustrative Case


The program used in this chapter to describe characteristics of successful formal

programs for preparing school-leaders on the high ground is one offered byOISE's Centre for Leadership Development. It is a program leading to certifica-

tion or eligibility to be considered for a school-administration position: principalor vice-principal. The program is offered in three sections. Part 1 and Part 2consist of approximately 125 hours of instruction each. The 'practicurn', whichtypically intervenes between Parts 1 and 2, involves school-based leadershipexperiences which take place in parallel with one's normal job. A special versionof the practicum awarding credit toward a graduate degree is also available.

This certificate program was first initiated in 1985. At the time of writingthis chapter, approximately 600 participants from three different sites in Ontariohad completed the program. Through a contract with the Department of Educa-tion in the Northwest Territories, an adapted-version program has also beenoffered to approximately 200 participants in that region of Canada since 1987.During the five-year period of the operation, each cycle of the program has beenevaluated both formatively and summatively. Components of this evaluationprocess have included pre-course surveys or other needs-assessment strategies,

assessment by candidates at regular intervals (usually weekly) throughout the pro-

gram, end-of-course 'suminative evaluations by candidates and staff participantobservation by a full-time course evaluator, and evaluation by a Ministry-of-

externpl monitoring team. While all five of these coin-

Developtng Expert Leadersl- p for Future Schools

A similar claim has also been made independently by Blum and ButLi- (1989),based on comparing the program with many others available elsewhere in NorthAmerica and Europe.

The Context in which The Case Program Was Designed and Implemented

All formal leadership programs are influenced by conditions or constraints intheir context. That is certainly the case for the program described in this chapter.Some understanding of those forces is required if the lessons learned from thisprogram are to be usefully applied in other venues.

The certification of principals is currently mandatory in only three regions ofCanada: New Brunswick, Ontario and, as of 1989, the North-West Territories.In the United States virtually all states report some form of principal certificationalthough there is considerable variety in the nature of qualifications necessary foraccreditation. In many North American states or provinces, prospective school-leaders are expected to have taken, or be willing to take, additional qualificationcourses which may or may not form part of a degree-granting program. Amer-ican state requirements usually include at least sonie graduate-level universitycredit courses in educational administration. Preparation programs in Canada andmany American states are delivered by faculties of education, at various sites,under contract with the local, provincial, or state education agency. Thus mostregions of Canada and the United States make available varying levels of pre-service training and in-service professional development. However, in manycases, these programs are not legislated requirements and may not result ingovernment certification.

Requirements for candidate entry into formal training programs can varywidely. In the United States there is an increasing trend towards the requirementfor a Master's degree. In fact the report of the National Policy Board forEducational Administration (1989) recommends that a Doctor-of-Education de-gree be the requisite for principal certification throughout the United States. InOntario, entry requirements include certification in three of the four divisions ofthe school program (primary, junior, intermediate, senior), program-specialistqualifications in two subject areas or completion of a Master's level graduatedegree. Candidates must also hold an Ontario teaching certificate and possess aminimum of five years teaching experience, two of which must be in Ontarioschools. The North-West Territories' (NWT) newly legislated program requiresthat course candidates hold a valid NWT teaching certificate, have at least a yearof teaching experience in the NWT, and must secure the recommendation oftheir regional superintendent for enrollment in the program.

In any given province or state there can be considerable variation in emph-asis among the objectives and methods of instruction across course sites. Forexample, some programs focus on a specific conception of the principal's role(e.g., instructional leader) or a generalized approach to carrying out that role(e.g.. school-improvement procedures). Other programs are structured by theNational Association of Secondary School Principals (NASSP) leadership-skillsmodel. Still other courses take a more traditional, thematic, 'issues dependent' orparticipant-controlled approach, sometimes termed the 'smorgasbord' approachby critics. Until recently such variations in approach were considered quite


1 9

Formal Programs tor Developing Expert School Leadership on the High Ground

acceptable For example, the Ontario Ministry of Education tolerates and evenencourages such variations in approach, providing that they are perceived asstemming from legitimate differences in regional contexts and/or sincere effortsto experiment with innovative course designs. However, the recent report of theNational Policy Board for Educational Administration (1989) illustrates an in-creasing consensus that reforms are required to establish a uniform set of stand-ards for the preparation of school administrators.

Principal-preparation programs in Ontario and the North-West Territoriesare typically delivered as two separate courses scheduled during the summermonths. As a result, candidates tend to become certified over a two-year period.In recent years, winter courses have been offered in some regions. These wintercourses operate on weekends and/or evenings between October and April. Anin-school practicum project is usually completed between courses during theschool year following completion of the Part 1 program. Summer courses are oftwo to four weeks in duration. Shorter courses, such as the ten-day NWTprogram or the three-week University of Western Ontario program, are veryintensive and require full-time residency. The longer, more relaxed four-weekformat employed by most Ontario faculties usually operates four days per weekand may or may not be residential. All Ontario courses, as well as the North-West Territories courses, make use of a variety of small and large group activ-ities. They also profess a sensitivity to the special needs of the adult learner.

Preparation courses in the United States arc more typically tied to graduate-degree prograni where candidates enroll in a set of credit courses. Considerablevariation in the number and nature of required courses exists among institutions.Similarly, field-based practicum projects are required in some states and not inothers. Typically, these are taken on a part-time basis rather than through asustained residency of a year or more, as has been recommended by the NationalPolicy Board for Educational Administration.

Instructors arid presenters in both the Ontario and North-Wcst Territoriesprograms available in Canada are, typically, practicing administrators. Ministrypersonnel, senior district administrators, trustees, social workers and other pro-fessionals often participate as presenters. University faculty play a significant role

in the design of these programs and frequently contribute to the implementationof selected components. Ministry or Department of Education personnel alsousually perform course-monitoring functions. The patterns arc similar in somestates of the United States, however, in other regions instruction is entirely in thehands of university faculty with minimal involvement of practitioners from thefield. This often Fives ris6 to questions of the relevancy of these programs:aspreparation for adniinistrative practice.

Program Goals: The Nature of Problem-Solving Expertise on the High Ground

As noted earlier, the coatent of many formal programs has been criticized forbeing unrelated to the achievement of valued outcomes in schools, not projecting

a coherent image of thc role, or not acknowledging the full range of demands onschool-leaders. The case program has avoided these limitations partly through its

research base.The program objectives of our pre-service program are derived from the


Developing Expert Leadership for Future Schools

research findings described in Chapter 5 and subsequent research specificallyfocused on school-improvement procedures. The former paints a multi-leveldescription of growth in high ground problem-solving expertise among school-leaders. By way of review, the four-component model of high ground problem-solving included:

Goals: the long-term, internalized aspirations held by school-leaderswhich form the basis for their decisions and actions; such goals might bederived from a school-leader's visionFactors: those aspects of the school, experienced directly by students,which significantly influence what they learn and which can be in-fluenced by school-leadersStrategies: clusters of related actions or procedures used by school-leaders to influence factorsDecision-making: processes used to identify and choose from alternativegoals, factors, and strategies

For a summary of the problem-solving processes associated with each componentof the model, the reader is referred back to Table 5.1 in Chapter 5. Thosequalities associated with the more expert levels of practice, as summarized inTable 5.1, served as general goals for Part 1 of our program, in particular.

Table 11.1 illustrates the type and specificity of objectives used to guide theprogram. In this case the objectives are based on the goals' dimension of the highground problem-solving model which is concerned with the nature, sources, anduses of school-leaders' goals.

While the four-component model of high ground problem-solving providesa framework for the certification program, as a whole, Part 2 of the programincorporates a set of knowledge and skills about school improvement. These areconceptualized within a five-phased school-improvement procedure (Leithwood,Fullan and Heald-Taylor, 1987), intended as a guide for school-leaders. Thephases of the procedure include:

Establishing the climate for changeIdentifying goals for school improvementSelecting and/or developing strategies or programs for achieving goalsImplementing selected strategies or programsInstitutionalizing selected strategies or programs

Each phase includes a set of general steps to be taken by the school-leader,identifies probable obstacles the school-leader might expect to encounter andprovides advice about how to respond to such obstacles. The features of eachphase have been distilled from a systematic review of recent school-improvementresearch. Traditional school-leader tasks such as program and teacher evaluation,relations with district staff and selection of curriculum materials arc treated aspart of the school-improvement process. Much of the knowledge and skillacquired during Part 1 of the program about 'Goals, Factors, Strategies andDecision-making' is applied as part of the school-improvement process, as well.In this sense, the major purpose of Part 1 is to give participants a clear image of



Formal Programs for Developing Expert School Leadership on the High Ground

Table 11 1 Objectives for Week One - July 4-7


Candidates will

1 Analyze the Ontario Goals of Education

2 Compare the Ontario Goals of Education with other goals proposed for public education

3 Identify potential problems or discrepancies among such goals and oblectives

4 Examine the principal's moral and ethical responsibilitiet, in relation to educational goals

5 Know how to derive a set of educational goals suitable for a school, given thedifferences in religious, cultural, and rawl contexts

what expert school-leaders do on the high ground; Part 2 helps show them how todo it.

Program Means: Justifying the Choice of Instructional Strategies

The delivery of many formal preparation programs has been criticized, as men-tioned earlier, for the use of generally weak or inappropriate forms of instruction.As a whole, such instruction is transmission-oriented: that is, based on an 'emptyvessel' model of the learner. Knowledge and skill are conveyed to the learnerwith limited interaction and too few opportunities for the learner to make theknowledge personally meaningful. In contrast, the case program has evolvedfrom one which had significant reliance on transmission instruction in its begin-nings, to one which is now highly transactional.

Both this chapter and Chapter 12 make explicit the conception of learningwhich underlies the formal programs described in those chapters. Then wedescribe the specific instructional strategies used in the programs and illustratehow those strategies arc incorporated in a slice of the life of each program.

The starting point for our conception of learning is the typc of knowledgethe programs are intended to develop. In the case of the program described inChapter 12, we call this 'useful, strategic' knowledge; in this chapter, we call this'declarative and procedural' knowledge. In fact, both programs share some simi-lar intents but different emphases with respect to types of knowledge. Moreemphasis is devoted in this chapter to declarative knowledge and more in Chap-ter 12 about useful, strategic knowledge. Useful strategic knowledge is alsoprocedural knowledge. The difference of consequence is that in this chapter weare concerned with how a school-leader acquires already developed and condifiedprocedures for solving routine problems. In Chapter 12, we arc concerned withhow school-leaders develop their own, often unique, procedures for solvingnon-routine problems. Such distinctions notwithstanding, these two chaptersdevelop one, not two, conceptions of learning. Differences between chaptersreflect the type of knowledge of greatest emphasis in the program being



Developing Expert Leadership for Future Schools

described. This chapter presents a relatively simple, skeletal model of learningappropriate to the development of all the types of knowledge addressed in bothchapters. Chapter 12 presents a relatively well developed extension of that basicmodel as it bears on the acquisition of useful, strategic knowledge.

The most direct source of thc learning model outlined here is contemporaryinformation-processing theory. Such theory admits to no definitive formulation.It is still rife with unresolved problems, as one would expect of any field ofstudy, subject to the amount of current work characteristic of this one. Neverthe-less, the brief synthesis provided here is generally consistent with more extensiveformulations that are to be found, for example, in Schuell (1986) and Calfee(1981). We present this brief summary of relevant information-processing theoryto explain our instructional methods. Its application within formal programs iscovered in the succeeding section entitled 'Instructional Strategies'. Those readersless interested in the theoretical underpinning may wish to skip to that section.

Acquisition of Declarative and Procedure Knowledge

Contemporary accounts of information processing stress the goal-oriented natureof human functioning and describe mental structures and processes associatedwith the resolution of problems standing in the way of goal achievement. Threestructures dominate this description: the Executive, Short-Term Memory (STM)

also called working memory, and Long-Term Memory (LTM). The Executiveis the primary location of both short and long-term goals (or aspirations). Onceperceived, information from the external environment is screened or assessed bythe Executive to determine its relevance for goal achievement. Informationjudged to be irrelevant is given no further attention; if judged to be potentiallyrelevant, information is passed on to STM. Beyond the limited processing spaceof STM or working memory and its capacity to integrate bits of information fortreatment as a single piece, little is known about the functioning of STM. Itspurpose, however, is to make sense of information passed on to it by theExecutive. It does this by searching through the virtually unlimited storage spacein LTM. Structurally, this space is represented as clusters or nodes of informa-tion, typically referred to as schemata, many of which are associated in networks,sometimes organized hierarchically. Relatively undemanding forms of sense-making take place when, through simple matching processes, STM locates ex-isting schemata or schematic networks capable of assimilating new information.More demanding forms of sense-making for instance, problem-solving - usual-ly demand modification of existing schemata or schematic networks to accom-modate novel aspects of information.

There is considerable debate about the nature of schemata. For presentpurposes, two distinct types are distinguished in LTM. 'Knowledge schemata'encompass facts, concepts, principles, and personal theories as well as affectivedispositions toward these elements. STM seeks out relevant schemata of this typein its attempts both to identify those elements or factors in the environmentwhich influence goal achievement and to determine the conditions within eachfactor that must be met if goals arc to be achieved. Such conditions having beendetermined. action may be required to meet them. Actions are guided by 'pro-cedural schemata': structurcs which indicate how to act and the steps to take.



Formal Progratns for Develop:lig Expert School Leadership on the' High Ground

Superordinate procedural schemata (sometimes called executive strategies) existto coordinate highly complex sets of actions.

Knowledge structures or schemata become increasingly sophisticated as theyare reorgan:. -I to incorporate additional pieces of related information and as theassociations among such schemata increase. Such sophistication is a function ofactive attempts to make meaningful more and newer information. And as newinformation is subsumed by existing knowledge schemata, the potential formeaningfully processing subsequent information increases. Actions become moreskillful (effective) as procedural schemata become potentially more adept inaccomplishing their ends, as overt beha-iors reflect more accurately the image ofskilled performance encapsulated in such schemata, and as the use of proceduralschemata becomes less conscious and more automatic. High levels of automatic-ny permit effective responses to environmental input without the need for pro-cessing such input through STM; this reduces response time and leaves theseverely limited information-processing space of STM available for handlingother problems.

Information-processing explanations of motivation begin with those inter-nalized goals located in the Executive. People are normally motivated to engagein behaviors that they believe will contribute to goal achievement. Strength ofmotivation to act depends on the importance attached to the goal in question andjudgment about its achievabihty. Motivational strength also depends on judg-ments about how successful a particular behavior will be in moving toward goalachievement (Bandura, 1977).

An information-processing view of the learning process is at the core of themodel of learning, underlying the prototype curriculum for school administra-tors. Nevertheless, there are two additional theoretical threads that serve not somuch to add to this view of learning as to emphasize and highlight several of itsfeatures. The first such thread is social-interaction theory (e.g., Simpson and(;albo, 1986). This theory stresses the dynamic nature of communication be-tween people in the creation of personal meaning. Because each of those involvedin communication actively brings different intellectual 'histories' to bear in theirattempts to construct such meaning, the outcomes of communication can neverbe entirely pre-determined. This has significant implications for the role ofinstructor as well as for the choice of instructional techniques, elaborated later inthis chapter. In brief, however, the implication is stated succinctly by Simpsonand Galbo (1986) (although they have a normal classroom context in mind):

The quality of a particular interaction is not entirely predictable, for theultimate form is determined by the participants at the time of encounter

llnstructors I must rely upon information gained through interactingwith students ... to determine some of the ultimate specifics of instruc-_tion. Some parts of the instructional process may be directed more bythe interaction of students and teacher than by the consciously deter-mined behavior of the teacher. The process is most useful in enhancing acarefully derived general lesson plan, especially in the hands of a superiorteacher with very clear objectives in mind. (pp. 45-50)

This position stresses the importance of interaction during learning - not onlybetween the learner and the formally designated instructor but with the learner'speers, as well.



Developtng Expert Leadership for Future Schnols

A second theoretical thread is adult-learning theory (e g , Merriam andCaffarella, 1991). While this is a highly derivative body of theory at present, itdoes provide a compelling argument for special attention to the greater accumu-lated contents of adult learners' long-term memory in comparison with that ofyounger learners. This reservoir of knowledge, skill, and affect is at once apotentially vast resource for sense-making and a relatively firm 'substance' tomodify and extend. Adult-learning theory also draws attention to:

the relatively complex and richly integrated organization of the contentsof long-term memorythe relatively high ego investment of adult learners in their past experi-ences and accomplishmentsthe relatively well established, clearly defined, personal goals the adultlearner brings to the educational experience

As did social-interaction theory, this position stresses the importance of interac-tion during learning. In addition. adult-learning theory supports instructionalstrategies that allow learners a significant role in shaping the nature and directionof their own instruction.

Instructional conditions which foster learniugWhile a model of learning, such as the one just outlined, is not synonymous witha model of instruction, its 'implications' for instruction are relatively obvious andsome have already been noted. In this section, nineteen such implications areassembled; they are conditions that can be met with an array of instructionalstrategies examined below. These conditions arc loosely associated with themental structures hypothesized by contemporary cognitive psychologists toaccount for mental functioning and, taken together, constitute what can bereferred to as 'transactional' instruction: instruction designed to be as personallymeaningful to the student as possible. Instruction is increasingly effective to theextent that it:

- The Executive -

Provides opportunities for the learner to clarify goals to self and to theinstructor

2 Demonstrates relevance of new information to the learner's internalizedgoals

- Short-Term Memory -

3 Helps the 'earner organize information into related 'chunks' for moreefficient processing

4 Introduces new information to the learner in small, manageable incre-ments

5 Provides the learner with immediate opportunities for making links tocontents of long-term memory



Formal Programs for Developing Expert School Leadership on the High Ground

Long-Term Memory (general)

6 Diagnoses contents of the learne.'s long-term memory relevant to use inmaking sense of new information

Long-Term Memory (declarative knowledge schemata)

7 Assists the learner in matching new information with as many existingknowledge schemata as possible

8 Assists the learner in expanding, modifying, or adapting existing sche-mata in order to make new information meaningful

9 Assists the learner in linking together previously independent schematain order to make new information meaningful

10 When no relevant knowledge is stored in long-term memory, assists thelearner to build new schemata and practice its retrieval

- LonpTerm Memory (procedural knowledge schemata) -

11 Provides the learner with initial procedural schemata by modeling,verbal description, and the like

11 Provides the learner with opportunities to act (perform) in accordancewith initial procedural schemata

13 Stimulates the learner to reflect on the discrepancies between his/herperformance and his/her procedural schemata

14 Provides the learner with feedback designed to increase the sophistica-tion of procedural schemata and reduce the discrepancy between per-formance and procedural schemata

15 Extends the learner's opportunities for practice with feedback untilperformance is sufficiently skillful

Motivation (multiple structures)

16 Clarifies for the learner the relationship between new information andhis/her own goals

17 Formulates the goals for learning in a sufficiently incremental way thatthe learner sees their achievement as feasible

18 Convincingly demonstrates the value of achieving the goals for learn-ing as contributing to achievement of the learner's own, internalizedgoals

19 Establishes a relationship between what is to be learned and storedknowledge and/or skill about which a person feels positively

Instruaional Strategies

In this section, we address six components of instructional strategy. These are:objective-referenced instruction, the role of instructors, techniques of instruction,program evaluation procedures, thc critical role of practicum, and selection ofmaterials.

200 177

Developing Expert Leadership for Future Schools

Objective-referenced instructionExperience with our program suggests that using specific defensible objectives tostructure program design and delivery allows instructional staff, as well as coursecandidates, to focus their efforts on the attainment of specific course outcomes.OISE programs, for example, are typically organized into four or five com-ponents, each of several days' duration, with clearly stated objectives keyed toassigned readings, small group activities and plenary presentations. One featurethat appears to be crucial to the success of the programs is that instructional staffand candidates regularly make reference to and use the course objectives.

Objectives are used in the program to reinforce and clarify the intendedfocus of the course, measure the progress of the course, evaluate the extent towhich objectives have been attained, and tie the diverse readings, discussions andactivities to a specific image of the principalship which candidates are intended tointernalize. This reliance on clearly articulated program objectives for programdevelopment and evaluation purposes was a distinctive characteristic of the ori-ginal OISE program in 1985. However, what has evolved since that time is anincreased dependence on the daily or weekly objectives by candidates and staff asa means for renewing the course focus as well as linking together the variouscomponents of the course experience into a consistent image of effective practice.

Instruaor rolesThe 'relevance' of programs for school-leaders hinges on the extent that courseobjectives reflect the knowledge, skills, and affect required for expert schoolleadership. It is possible, however, for the substance of a program to be 'relevant'yet for the program still to suffer from lack of utility. Utility is a function of notonly program objectives but also the nature of instruction designed to achievethose objectives.

Often, criticism of program utility is expressed in terms of excessive atten-tion to theory and not enough concern for practice. More precisely, however,utility means 'capable of being put to use'. And while it is true that bad theorycannot be put to use with much advantage, good theory has great potential utilitythrough its power to predict and control. This is especially so in otherwise highlyuncertain environments like those inhabited by school-leaders. Indeed, manyschool-leaders have developed quite elaborate, although often implicit, 'theoriesin use' (as Argyris, 1982, would call them) to guide their work. Unpacking themeaning of utility in this way raises the question: What is 'good' theory from theperspective of the school-leader's job? How can the job be done more expertly?What can be done to ensure that a program as implemented reflects these featuresof good theory and is, thus, useful?

Part of the answer to this question about good theory is to be found in threefeatures of some of those 'theories in use' already guiding many school-leaders.Such theories, first of all, are theories for action; they are designed for thepurpose of prescribing what ought to be done in response to some problem. Byfar the bulk of current, formal administrative theory has as its purpose descrip-tion and explanation; this is a sometimes helpful but never sufficient basis foraction. The objectives of our program for thc high ground, in contrast, are basedon theories for action, ones which conceptualize how school-leaders can beeffective in bringing about school improvement. Sccond, 'theories in use' haveusually been subjected to considerable empirical verification, albeit a highly



Formal Programs for Developing Expert School Leadership on the High Ground

personal, unsystematic form of verification. While there is a long history ofprescriptive theory in administration (theories for action), its empirical verifica-tion is woefully limited.

One response to the importance of basing a school-leader program onverified theory has already been discussed; the bulk of our program's objectivesfor the high ground remain well rooted in empirical research data. This ground-ing, however, was necessarily limited to the sites in which data were collected.Extensive use of practicing school administrators as program instructors is anadditional form of verification. These instructors are able to relate the generalizedtheory for school-leader action to their ov.n work and convey its utility, throughinteraction with participants, from that perspective. They are also able to supple-ment the generalized theory, when interaction suggests that is necessary, fromthe stock of their own professional experience.

The same response, having practicing school administrators as instructorsand co-developers of the program, is a way of recognizing a third powerfulfeature of school-leaders' 'theories in use'. Many school-leaders"theories in use'are sufficiently operationalized so that their implications for application to specificproblems in each school-leader's own school context are extremely clear, at leastto the holder of the theory. In contrast, much formal administrative theory isremote from specific action and often ambiguous in the guidance it provides foraction in a particular context. Instructors need to be able to add specificity to thegeneral theory for action guiding the course and help students make meaningfulapplications to their own context.

The sele-ction of appropriate instructional staff is an especially critical factorfor ensuring the effectiveness of a formal preparation coursr. A number ofinsights have been gained by OISE personnel in these matters as a result of fiveyears experience. The selection of the course principal or coordinator is particu-larly critical. A course principal must be willing to do more than just manage thecourse. He or she must be knowledgeable about, and committed to, the concep-tual framework of the course and take steps to ensure that it is honored. Thecourse principal, as well as group leaders, must model effective practice and,when necessary, do whatever is required to guarantee the integrity of the course.All group leaders should ideally be experienced practitioners, although this isoften not the case in formal training programs offered in a university setting. Thejustification for this assertion is the aspirants' need for appropriate modeling.Beyond this important requirement, course staff should be representative in anumber of ways. Group leaders should be highly experienced principals,although one or two relative newcomers to the principalship may also be desir-able in the interest of providing a fresh perspective. Gender, racial and religiousfactors should also be balanced. Above all, group leaders must understand thattheir role is chiefly to facilitate group processes, to act as a resource whenrequired, and to model effective practice. They are not there to dominate discus-sion, pass along war stories or launch an independent course process.

There is another side to the utility dilemma which cannot be adequatelyaddressed, simply by modeling the features of school-leaders' implicit theories.While many educational programs have been accused of being `too theoretical',others have been described as trivial and mundane (the most common criticismleveled at principal-certification programs in Ontario prior to about 1980). Thiscriticism implies that there is an excessive focus on individual school-leader's



Developing Expert Leadership for Pi, . -chools

espoused theories (Argyris. 1982). TE value to others of such espoused theoriesdepends on their congruence with 'thi...)ries in use'. Espoused theories which donot closely capture 'theories in use' do not benefit from the empirical verificationthey are normally associated with.

The value of individual school-lcader's espoused theories to others alsodepends on their external validity. Espoused theories of no demonstrable effect inmultiple school contexts are probably of interest only to the espousers and theirimmediate families. The process used in both designing and implementingour program attempted not only to recognize, but actively to foster productivetension between theory and practice. Researcher-participants were forced toclarify the meaning of their research in specific cases and contexts. Instructorswere forced to examine the relationship between their 'theories in use' (andespoused theories) and the general theory for school-leader action, reflected inresearch-based descriptions of expert problem-solving.

Instructional techniquesA review (e.g., Sparks, 1983, Silver and Moyle, 1984, Hutson, 1981, Daresh andLaPlant, 1984) of pre-service and in-service education programs for teachers andadministrators generated a significant number of promising instructional techni-ques. Thesc techniques, which are potentially available to meet the learningconditions identified earlier, are as follows:

(a) Opportunities forlearners to identify someof their own needs andto participate in someprogram planning

(b) Lectures (givinginformation)

(c) Private reading andreflection

(d) Independent study(e) Demonstration of skills

by 'experts' (live, video)(f) Opportunities for

practice and feedback(coaching)



Role playingGuided group-discussionCase analysesSimulated casc problem-solvingSite visitsParticipant presentationsOpportunities for subgroupleadershipProvision of individualdiagnosis and counselingClarification and extension ofideas with peers throughdiscussion

Table 11.2 indicates which of these techniques seem most suitable in meetingeach of the nineteen learning conditions discussed earlier. Sonic of the attributesin Table 11.2 are speculative and might change, depending on more specificinformation concerning how the instructional technique is to bc applied; a num-ber of these techniques could be used with widely varying consequences.

Using the course process itself as a simulation of school-leadership activitieshas proven to be a particularly useful technique to minimizc the 'training foractivity' trap previously described in this chapter. Such an approach mightinvolve the rotation of the small group chairperson's role on a daily basiswhereby one person, or occasionally two working collaboratively, manage thegroup's affairs for a day. They may be required to chair group discussions,



Formal Programs for Developing Expert School Leadership on the High Ground

Table 11 2: Relating instructional techniques to learning conditions

Learning Condition

The Executive1 Clarify goals to self and instructor2 Demonstrate relevance of new information to internalized


Short-Term Memory3. Organize information into related chunks4 Introduce new information in small manageable increments5 Immediate opportunities for making links to contents of long-

term memory

Long-Term Memory (general)6 Identify contents of long-term memory relevant to making

sense of new information

Long Term Memory (declarative knowledge schemata!7 Matching new information with as many existing knowledge

schemata as possible8 Expand, modify, or adapt existing schemata in order to make

new information meaningful9 Linking together previously independent schemata10 Build new schemata and practice its retrieval

Long-Term Memory (procedural knowledge schemata)11 Provide initial procedural schemata by modeling and verbal

description12 Provide opportunities to act in accordance with initial

procedural schemata13 Stimulate reflection on discrepancies between performance

and procedural schemata14 Feedback to increase sophistication of procedural schemata

and reduce discrepancy15 Practice with feedback until performance is sufficiently


Motivation (multiple structures)16 Clarify relationship between new information and goals17 Formulate goals incrementally to foster perceptions of

feasible attainment18 Demonstrates value of learning to achievement of

internalized goals19 Establish relationship between learning and valued prior

knowledge and/or skill


a, c, h, na, h, n

b, c, d, 1, ob, c, e, g,f, g, h, 1, o

c, d

a, c, d, I. o

a, c, d, e, f, g, 1, m

c, d, f, 9.f, g, 1, 1, m, 0

b, e, k

f. g, I, m

f, g, I. n

f, g. I. m, n

I. m, n

a, n, oe, f, g, I. m, n

a, c, m

a, n

schedule activities, manage time and solve problems as they occur. In thesesettings, the group group-leader becomes a resource rather than a director ofactivity. Course instructional staff can further model effective practices by openlydescribing their perceptions and responses to real problems which crop up duringthe course. Candidates quickly catch on that the course process is an ideal testingground for their newly-learned management and leadership skills.

Although the objectives of a formal preparation program resulting in accre-ditation should be relatively non-Legotiable once the course design has been


Developing Lxpert Leadership _tor Future Schools

validated, candidates can s:ill be given considerable freedom, particularly in smallgroup sessions, to make choices about what and how they will learn, and howthey will use their time. Indeed, it is arguable that more alternate activities shouldbe made available than time would permit addressing. This forces candidates toparticipate in interesting simulations of the priority setting and time-managementexercises principals frequently encounter in real schools.

Through a candidate-committee system established as part of the courseprocess, participants soon recognize that control of the course experience isshared with them to a considerable extent. For example, in OISE programs, fourcommittees are typically established: program, communications, evaluation, andsocial. A final strategy frequently employed in formal preparation programsinvolves encouraging candidates to share any special expertise they may possessthrough candidate-organized workshops. These workshops can be a regular partof the course which is usually scheduled towards the end of the program. Thisstrategy has been found to bc particularly useful when delivering courses inisolated areas (e.g., the North-West Territories of Canada) where the inclusion ofcultural issues is a critical requirement and the availability of expert presenters canbe limited.

Most of these techniques are probably better thought of as: generalapproaches to instruction that can be further developed once one is clear aboutthe objectives to be met, some of the preferred learning styles of students, theamount of variety required to maintain energy and interest over the entire periodof the program, and the skills and preferences of instructors (although theselection of instructors should be (lone so as to avoid restricted choices oftechniques, for this reason).

Evaluanon proteduresFive or more components can be recommended for inclusion in the evaluationprocedures for a state-of-the-art leadership-development program. These mayinclude pre-course surveys or other needs assessment instruments, assessment bycandidates at regular intervals throughout the course, end-of-course 'sununative'evaluations by candidates and staff, participant observation by an objective courseevaluator, and Ministry-monitoring team reports.

An important consideration concerning evaluation of pre-service programshas to do with the purpose of evaluation. Course-delivery personnel typicallymaintain a strong interest in two predominant purposes for course evaluation.First, the issue of accountability, particularly given contractual arrangementswith the departments or ministries of education is a primary concern. For themost part. end-of-course 'summative' evaluations and ministry or departmentmonitoring have been successful in meeting such accountability demands. Instru-ments which provide course-delivery personnel with 'satisfaction-lever data con-cerning various program components (e.g., meeting of objectives, readings,plenary sessions) are another example. These data derive primarily from re-sponses to rating scales and associated comments. At the completion of eachcourse, a summary can be made of responses to these items. The summary isuseful not only from the standpoint of meeting accountability needs but inproviding information for the second major purpose of course evaluation.

The second major focus, one which has been more central to ongoingoperations, is 'formative' evaluation or evaluation for course improvement. The



Formal Programs for Developing Expert School Leadership on the High Ground

needs assessment practices referred to earlier have served to sensitize coursecoordinators to the specific needs of a particular client group. Program modifica-tions based on these data can be readily made. Participant observation and dailycourse-evaluation data and comments on end-of-course instruments have alsoproven to be useful for modifying programs. Such instruments can provideprogram-delivery personnel with a wide range of comments and suggestions, thecareful analysis of which, in view of participant observation data, can lead tosome rather immediate, and responsive, mid-course changes. Suggestions associ-ated with end-of-course evaluation instruments can also help course coordinatorsprepare and plan for upcoming programs.

The evaluation process allows for another major use of formative data. Thatpurpose is to promote among candidates 'collective reflection' about things thathave been learned, thoughts that have been generated, and perceptions that havebeen shared. For example, in OISE-sponsored courses, an evaluation committeeis typically formed to assume responsibility for evaluation tasks. A major func-tion of the committee is to analyze thc evaluation data, but another function is tofeed that information back to the group. This has typically been done in the formof a short verbal summary by a committee member, prior to having candidatesengage in the next day's activities. This type of feedback tends to be very pos-itively received by canciisizzcs. It helps them to consolidate information byobserving how their peers have perceived it. Such information is viewed as veryrelevant by participants, especially those who have limited leadership experiencefrom which to draw (Cousins and Leithwood, 1986). Finally, this process alsocarries with it the side benefit of motivating candidates to complete the dailyevaluation forms.

A continuing problem with the evaluation of pre-service courses relates wellto Gaines-Robinson and Robinson's (1989) notion of the 'training for activity'trap. How does one measure the extent to which pre-service training impacts onperformance (current or future) in the schools? In a recent in-service program,OISE personnel used a more elaborate instrument for collecting end-of-coursedata. The instrument not only asked for the usual 'satisfaction with coursecomponents' data and suggestions for course improvement, but also posed ques-tions about the school-level impact of the course experience.

These data, albeic self-report and limited for that reason, permit a statisticalassessment of the relationship between course processes and course outcomes.The applicability to pre-service delivery of this approach is limited, due to thenature of the client. For the most part, candidates are not currently in leadershiproles and would be unable to assess impact at the school level as a consequence.Nonetheless, linking evaluation data to field-based practice is a goal worthy ofpursuing in the interest of meeting both accountability and course-improvementdemands.

practicurnSome aspect of formal training programs should be close to or in schools and theprogram should normally be sustained over a relatively long period of time.Unfortunately, virtually all Canadian and American pre-service courses arc con-ducted away from the candidates' normal work environment. This pattern isperhaps a matter of academic tradition unlikely to change for at least two reasons.First, centralized delivery may afford ministrics (departments) substantial



Developing Expert Leadership for Future Schools

opportunities to monitor and 1,1aintain a level of control over from and content.Second, lumping people together in central locations for short intensive experi-ences is probably fiscally and pedagogically more efficient. Most school districtshave neither the desire nor the resource capacity to run their own programs.

School-based practicum projects, as employed in Ontario and the North-West Territories, for example, begin to attend to this issue in some respects.However, while they provide opportunities for in-school application of formalcourse learning, the quality of the practicum experiences has been found to varygreatly depending on the suitability of the project, the commitment of thecandidate and supervisor to the exercise, and the appropriateness of the localschool setting for such a project.

The practicum experience scheduled between Parts 1 and 2 of the OISEcertification program is intended to accomplish three general purposes. First, itaims to assist participants in applying and refining the decision-making andproblem-solving schemata required for the effective use of developed proceduresin the contingent world of their own schools. This purpose is accomplished tothe extent that the practicum instructor (e.g., an experienced principal) models,provides opportunities for practice, and gives feedback to the participants abouttheir decisions and decisioa processes. Second, the practicum instructor mayfurther develop the procedural knowledge that participants bring to the practi-cum situation and extend the repertoire of procedures possessed by them. Boththese purposes are possible the practicum to the extent that effective proce-dures and problem-solving processes can be made relatively explicit.

The third purpose, while more difficult to accomplish, is a traditional ex-pectation for the practice of practicum experiences in other types of education.Such experiences include the clinical work of medical interns under the guidanceof a senior practitioner, the articling experience of the novice lawycr, and appren-ticeship activities associated with skill development in the fine arts and in craft-based vocations. These practices share in common the goal of acquiring the moretacit components of the experienced practitioner's repertoire of knowledge andskill those components that permit such practitioners to dcal effectively withproblems 'in the swamp' or 'situations of u certainty, uniqueness and conflict'(Schön, 1983, p. 16). Chapter 12 provides a more complete analysis of what isrequired to develop expertise in the swamp.

Schön (1987) accounts for the 'artistic aspects of practice' by reference to thistacit knowledge and skill. He suggests that these aspects of practice are acquiredin professional studios and conservatories by creating certain necessary conditionswhich he identifies as the:

freedom to learn by doing in a setting low in risk, with access to coacheswho initiate students into the traditions of the calling and help them bythe right kind of telling to see on their own behalf and in their own waywhat they need most to see. (p. 17)

What does this view of a practicum mean in the context of our certi-fication program? To accomplish the three general purposes discussedabove, the practicurn attempts to:


Provide a significant portion of time to observe experienced principals atwork on a range of non-trivial problems and to discuss with principals


Formal Programs for Developing Expert School Leadership on the High Ground

their intentions, how Lhey link their overt actions to those intentions,and how they manage their problemsProvide an opportunity for the participant to become involved in anadministrative problem on the high ground, with close coaching from auexperienced principal. This coaching is intended to foster what Schön(1983) refers to as 'reflection in action' in order to refine proceduralknowledge and problem-solving processesProvide an opportunity for the participant to become involved, with anexperienced principal, in an administrative problem in the swamp. Fre-quent discussion of the process used by the experienced principal andimplications for solving other problems should foster 'reflection inaction'

Selection of appropriate course materialsOur experience with the OISE program also recommends special attention to theselection of course materials. In Ontario and some parts of the United States,many candidates typically have graduate degrees which were previously earned.These individuals are likely to have a relatively high capacity for academicreading as compared to course participants in other regions without graduateschool experience. For the latter group, substitute readings from popular profes-sional journals can provide a good alternative. Alternatively, when difficultjournal articles or research reports must be used, the inclusion of a reading-guidecover page detailing and highlighting important points is a good strategy. Smallgroup discussion of important readings in a seminar setting is also good practicein that candidates are able to increase comprehension through discussion anddebate. Course duration is another factor which influences the amount of readingthat can be realistically assigned. There is little point in assigning readings thatcandidates will not or cannot read. Attention to preferred learning styles and amulti-media approach to the presentation of material is a preferred strategy;increased focus in this direction has characterized recent OISE programs.

Summary and Implications

In this chapter, we identified characteristics of formal programs for developingexpert school-leaders on the high ground. We used five years of experience anddata collection in the delivery of a school-leader certification program in Ontario.

A number of issues central to high-quality programs have arisen throughoutthis description. Nine of these are selected as especially in need of emphasis, or asnoteworthy implications for others designing formal programs for developinghigh-ground expertise.

1 A Coherent Research-based Image of School Leadership

Several problems may be overcome if the program builds from a research-basedconceptual framework; candidates are able to develop a comprehensive concep-tion of the role of school-leader; segments of the course may be tied together in acoherent pattern; the relevance of course objectives becomes apparent; individualissues are viewed in the context of the large: picture of the role; candidates will



Developm Expert Leadership for Future Schools

develop a sense that the validity of key issucs and practices is justified byresearch; and finally, unanticipated objectives determined or contributed bycourse participants are more likely to be congruent with the goals of the course.Course instructors should be knowledgeable about, and committed to, the par-ticula,.. framework used. Provision should be made to review the frameworkperiodically and to update it, based on emerging research.

2 Systematically Assess Client Needs

Knowledge and skills that are demonstrably relevant to clients' needs from theclients' point of view are likely to be much more useful. Within the confines of aconceptual framework, program implementors should systematically collect in-formation prior to course delivery and should use that information to modify theprogram. Variation in stages of development in candidates' leadership knowledgeand skills is likely to be enormous at the pre-service level. Instructors should beprepared to assess and respond to such variation. Data from previous versions ofthe program, participant observations, and such programs and informal contactscan serve to refine more formally collected needs' assessment data.

3 Maintain Objectives-based Focus

Candidates should regularly be made aware of the purposes of the program, andwithin bounds, have an opportunity to shape such purposes. Program objectivesshould be linked directly to the underlying conceptual framework but should besufficiently flexible to allow for continued review and adaptation as the needarises. Where possible, variation in course objectives should overlap with candi-dates' stage of development.

4 Adhere to Principles qf Ejecting' Instruction

Appreciation of different learning styles, transactional curriculum delivery,multi-media presentation materials, and continual involvement of course partici-pants in course delivery are important features of effective pre-service delivery.Candidates have little tolerance for being lectured to extensively, although, inrecognition of the need for content input, an appropriate mix of instructionalstrategies is recommended. Candidates find extremely valuable opportunities toshare with peers relevant experiences and to learn how othcrs would handleparticular problem situations. Learning materials should be sensitive to the candi-dates' stage of development and they should be provided with a healthy amountof 'hands-on' material.

5 Simulate Role Responsibilities

As part of the course-delivery process, candidates should be actively involvedin chairing small group discussions, organizing workshops, assuming course



I-ormal Program> for Developing Expert School Leadership on the High Ground

administrative responsibilities, and otherwise engaging in leadership functions.Activities designed to simulate school-leaders' responsibilities can enhance thedevelopment of procedural knowledge and minimize the apparently abstractnature of some program topics. This is particularly an importantconsideration whereleadership expertise is limited and candidates do not have extensive experiencefrom which to draw.

6 Include Practicum

Candidates typically work in environments where high-quality, supervised prac-tica can be readily undertaken. It is important to practice and reinforce program!earnings in on-the-job experiences. This type of activity provides candidateswith opportunities to refine procedural knowledge and to reflect on, and analyze,such experiences. Program-delivery personnel should be sensitive to candidates'motivation to 'fast-track' and take steps to ensure that practical experiences areadequately reflected upon.

7 Promote Social and Professional Interaction

Candidates probably learn as much from interacting and participating with oneanother as they do from course readings and presenters. Opportunities forinteraction with one another in either structured (e.g., home group) or un-structured (e.g., social occasion) situations should be provided regularly. Theseopportunities not only enable candidates to gain valuable insight from the experi-ence of others but may serve to consolidate information in their own minds.

8 Use Evaluation Processes for Multiple Purposes

Valuable insights concerning course development can be gleaned from evaluationconducted for accountability reasons. The sophisticated usc itself of formativeevaluation to improve course delivery can meet accountability needs. But evalua-tion can be used for at least one purpose other than accountability or develop-ment: it can bc used to stimulate reflection among candidates. Thinking aboutone's own perceptions and feelings about a program experience is certainly oneway to stimulate reflection. Thoughtfully completing a daily evaluation task;becoming involved in analyzing and reporting evaluation data to one's peers; andreceiving daily evaluation reports (either written or verbal) are all processeswhich arc likely to stimulate, among candidates, reflection about importantissues.

9 Work 'roward Assessinq Cource Impact

Many course evaluations arc too focused on process and pay little, if any,attention to outcome. Course developers would do well do give serious consid-eration to defining real course outcomes (e.g., other than course satisfaction) and



1.4 -J.

Developing Expert Leadersh,o for Future Schools

in some way attempting tu measure these The very nature of pre-service de-livery, in.smuch as most candidates are not in positions to implement adminis-trative strategies learned, stands as an obstacle to collecting data on impact.Nonetheless, this is an obstacle with which it is worth struggling. For thepresent, even in state-of-the-art pre-service programs, the value of particularpractices espoused by the program are justified by research findings on effectiveadministrative practices rather than by empirical evideir7e of improved perform-ance in the role of program candidates.


This chapter reviewed current practice in the delivery of pre-service programs foraspiring school administrators in Canada and, to a lesser extent, the UnitedStates. A critical analysis of traditional practices was also presented. Based on fiveyears of experience in the delivery of principal pre-service training programs andextensive evaluation data accumulated over the years, the evolution of OISEprograms and how developers have attempted to meet the challenges of pre-service delivery were discussed as an illustration of a formal program for de-veloping effective leadership on the high ground.

Traditional programs were described as being predominantly 'issues-oriented' and lacking in coherence and an adequate conception of the school-administrator role. The case was made for structuring pre-service deliveryaccording to a research-based conceptual framework. Key implementation andevaluation elements of the OISE program were described as delivery strategiesthat meet the developmental needs of candidates, encourage reflection amongcandidates as courses proceed, address ill-defined or swampy role responsibilitiesthat principals face daily, and attempt to 'train for outcome' as opposed to coursesatisfaction. Consideration was given to problems that emerge when course-delivery personnel attempt to implement these strategies.

The chapter concluded with a set of recommendations based on what theauthors believe to be principles of effective pre-service delivery practice. Theappropriateness of individual recommendations and the weight each carries willvary from one preparation program to the next. Nonetheless, serious attempts toadhere to them seem likely to result in improved pre-service programming.


Chap ter 12

Characteristics of Formal Programsfor Developing Expert SchoolLeadership in the Swamp

As Chapter 11 indicated, there is a relatively large body of well examinedexperience on which to draw in determining the characteristics of formal prog-rams that will be successful in preparing future leaders for the high ground.Much less experience has been had in preparing school-leaders for the swamp.Nevertheless, that experience has been instructive with regard to program char-acteristics and gives rise to optimism about the possibility of improving expertisein the swamp.

The program characteristics described in this chapter are based most directlyon research related to versions of a four to five-day program we have designed,implemented and refined through four iterations over a three-year period. Severaldifferent types of data were collected each time the program was implemented.The most comprehensive research was undertaken in conjunction with the firstiteration of the program. On this occasion, the program was implemented withtwenty-two school-leaders (fourteen principals, eight first-year vice-principals)and changes in their expertise were compared with a matched group of sixteenprincipals and vice-principals. Data collected before and after the program fromboth groups included the participants' own appraisals of changes in their exper-tise, and written solutions to case problems which were analyzed in detail by theresearchers and holistically by two expert school-leaders working independentlyof one another. The same types of data were collected from the nincty experi-enced school-leaders in the second iteration of the program but there was nomatched control group included in this study. Only participants' self-appraisalsof change were collected for the third and fourth iterations.

Each time the program has been implemented, three sets of questions haveserved as the focus for data collection. First, can school-leaders' expertise insolving swampy problems be improved? Second, will an instructional focus ongeneral problem-solving processes used by expert school-leaders enhance theproblem-solving capacities of others? Finally, what are the most promising prog-ram characteristics for improving school-leader expertise in the swamp? This lastquestion is the main focus of this chapter. Nevertheless, unless thc first twoquestions can be answered affirmatively, it is a meaningless question. Our data,in line with evidence provided through research in other fields (e.g., Nickerson,1988-89) strongly suggest such affirmative answers. In this chapter. we do not



Developing Expert Leadership for Future Schools

provide a detailed report of the results of our data analysis. For those wishing toreview these data see Leithwood and Steinbach (1991).

The remainder of the chapter describes the central characteristics of a formalprogram for developing cxpert school leadership in the swamp. These character-istics are of two types: characteristics related to the purposes or goals of theprogram and characteristics of the instructional strategics used in the program. Ineach case, the theoretical justification for each characteristic is described in somedetail. Some readers may consider this unnecessary for them and may wish toread only the conclusions of our theoretical musings, the `practical' stuff. In fact,we believe these theoretical musings to be highly relevant for anyone interestedin their own development or in assisting with the development of others. Ourgoal is to insert a 'light bulb' into the 'black box' of the school-leader's mind.Being able to see better how the insides of the `black box' work increases ourflexibility and effectiveness as instructors by reducing our dependency on instruc-tional prescriptions from others. It also allows us to exercise more productivecontrol over our own learning.

Program Goals: The Nature of Problem-Solving Expertise in theSwamp

Expert Processes Identified in the Research

Problem-solving processes which our experimental program aimed to developemerged most directly from research partly described in Chapters 6 and 7. Asnoted there, this research has: examined school-leaders' problem classificationand management processes (Leithwood and Stager, 1986, Leithwood, 1988);developed a grounded model of school-leader problem-solving alone (Leithwoodand Stager, 1989) and in groups (Leithwood arid Steinbach, in press b); examinedthe flexibility of school-leaders' cognitions (Stager and Leithwood, 1989); andinquired about the role of values in problem-solving (Begley and Leithwood,1989, in press, Campbell, 1988). Differences in problem-solving processes due todifferences in role or organizational context (Leithwood and Steinbach, 1989)have been examined. As readers of Chapters 6 and 7 will know, much of thisresearch has examined differences in the problem-solving processes of `expert'and `non-expert' school-leaders. As in the research of many others, using expertvs. novice designs (reviewed by Alexander and Judy, 1988) our long-term pur-pose has been to clarify those aspects of expertise that might become a productivefocus for selection, evaluation, and professional development of principals.

By way of quick review, a central result of this research has been a six-component model of problem-solving including:


Interpretation: principals' understanding of the specific nature of theproblem, often in situations where multiple problems may be identifiedGoals: the relatively immediate purposes that the principals are attemp-ting to achieve in response to their interpretation of the problemPrinciples/Values: the relatively long-term purposes, operating princi-ples, fundamental laws, doctrines, values, and assumptions guiding theprincipals' thinking


Formal Programs for Developing Expert School Leadership in the Swamp

Constraints: 'barriers or obstacles' which must be overcome, if anacceptable solution to the problem is to be foundSolutior Processes: what the principals do to solve a problem in lightof their interpretation of the problem, principles, goals to be achieved,and constraints to be accommodatedAffect: the feelings, mood and sense of self-confidence the principalsexperience when involved in problcm-solving

For a summary of the problem-solving processes associated with each com-ponent of the model, the reader is referred back to Table 6.1 in Chapter 6. Thoseprocesses associated with experts summarized in Table 6.1 served as general goalsfor the instructional program. As is evident in Table 6.1 and discussed extensive-ly in Chapter 7, values play an integral role in problem-solving as do the problemsolver's moods or affective states through their influence on problem-solvingflexibility. While both of these components of problem-solving merit instruction-al attention, our experimental program touched on them only lightly. This wasbicause we were not confident in the instructional approaches available to us.Instruction was limited to making participants aware of research results concern-ing the nature and role of values and mood in the problcm-solving of experts.The program also encouraged self-instruction and self-evaluation as a means ofregulating mood, techniques which sPeni to be helpful for this purpose (Meichen-baum, 1977, Wine, 1971).

Justifying a Focus on General Processes

Greatest emphasis in the program on which this chapter is based was focused onthe other components of the model, which are a set of general cognitive skills andproblem-solving strategies (e.g., problem interpretation, goal setting, identifyingconstraints, and solution processes). The focus is important to acknowLdgebecause of the long standing, continuing debate about the relative contribution toexpertise. in many fields, of domain-specific (or local) knowledge (e.g., know-ledge about teacher evaluation) as compared with generalizable, content-independent skills and strategies (e.g., Ogilvie and Steinbach, 1988).

Evidence brought to bear on either side of this debate is reasonably compell-ing: on the 'domain-specific knowlege' side, see Chi, Glaser and Farr (1985) andLesgold (1984), for example; examples of evidence on the 'general skills side'include Brown and DeLoache (1978) and Nisbett and Ross (1980). As Nickerson(198-89) points out, however, this controversy is primarily a question of emph-asis with widespread acknowledgment, in recent research, that both generalthinking skills and domain-specific knowledge are important. In the absence ofdomain-specific knowledge, one has nothing to think about. In the abseme ofreasonably well developed general thinking skills, one's knowledge may wellremain inert: that is, not be applied in circumstances where it has potential use.For instructional purposes, it seems reasonable to approach the matter in aconditional way. The probability that a school-leader will successfully solve aproblem is a function of both the availability of problem-relevant knowledge andgeneral thinking skills or heuristics and so both arc addressed in the program.

2 1.1A


Developing Expert Leadership for Future Schools

The primary focus of insti uction is on general problem-solving strategies withdomain-specific knowledge provided by colleagues in the course of solvingreal-life administrative problems.

A general strategies focus was chosen for three related reasons. First, expertpractice jn many professions is centrally concerned with solving ill-structuredproblems for which there is relatively little available content knowledge and noreadily available solution. As Schön (1983) suggests and as we discussed inChapter 6, it is problems 'in the swamp' that are of greatest interest to experi-enced practitioners in many professions, and there is a corresponding need tolearn how to respond to such problems in an effective manner. Evidence re-viewed in Chapter 4 sugg...;ts that as many as one in five problems faced byschool administrators arc ill-structured from their point of view. Second, wcbelieve that experienced school-leaders at the present time, contrary to Perkins'(1985) assertion, are likely to have more opportunity to acquire domain-specificknowledge than general problem-solving skills. Descriptions of the content ofcontemporary preparation programs provide one source of evidence in support ofthis belief (Leithwood, Rutherford and Van der Vegt, 1987, Blum and Butler,1989). Hence, improvements in present school-leaders' practices seemed likely byimproving their capacities to use their existing knowledge more effectivelythrough increasing their problem-solving skills. Finally, since our researchshowed that expert practitioners could be differentiated from their less expertcolleagues on the basis of their problem-solving skills, it seemed appropriate to atleast try to teach those expert processes.

We believe our choice of focus on general problem-solving skills for theexperimental program is justified by the case outlined above. Nevertheless, theterm 'general thinking or problem-solving skills' conveys an incomplete impress-ion of the knowledge we attempted to develop through the program. A morecomprehensive impression is conveyed by the term 'useful, strategic knowledge'.

Interpreting General Processes as Usefid, Strategic Knowledge

In the case of the experimental program, 'useful, strategic knowledge' refers tothe explicit strategies and heuristics associated in our research with expertise, aswell as the (usually) tacit knowledge required for its actual use in real-life schoolleadership contexts. Our mcaning of 'useful, strategic knowledge' is essentiallythe same as Sternberg and Caruso's (1985, p. 134) definition of practical know-ledge: 'procedural knowledge that is useful in one's everyday life'. Such know-ledge is strategic or procedural in the sense that it is knowledge concerned withhow to solve problems (e.g., how to clarify school goals) rather than knowledgeabout problem-solving (declarative knowledge).

The strategic knowledge of concern to us also had to be useful in two senses.First such knowledge had to be sufficiently detailed so as to be of direct use in thecontext of school-leaders' practices. This meant that the explicit problem-solvingstrategies identified in our research had to be combined with an extensive body of'ordinary' (1,indblom and Cohen, 1979) knowledge, usually tacit, in the mind ofthe learner; the importance of this type of knowledge is often overlooked indiscussions of expertise. This combination acknowledges both the limited cogni-tive guidance provided by explicitly described, general strategies and the con-



Formal Programs for Developing Expert School Leadership in the Swamp

ditional nature of practical krowledge (e.g., knowledge that: 'although yourteachers need more time to plan, Harry Smith doesn't want to be out of his classduring school hours any more than he is already'). With respect to strategicknowledge, expert practice requires explicit general strategies, detailed know-ledge about how to use them, and knowledge of the circumstances under whichtheir application is appropriate. For our purposes, 'useful' also meant that theknowledge provided by the experimental program had: to be relevant to solvinga wide range of school-leaders' problems, to contribute significantly to theirsuccess in solving problems, and to be potentially teachable. Perkins and Salo-mon (1988) argue that four conditions must be met for knowledge to be useful inthis sense. The research from which this program emerged (see Chapter 6)indicates that the program does meet them. That is, the program's focalstrategies: appeared to be used by experts in solving problems; played an impor-tant role in problem-solving; were transferable across many problems; and werecommonly absent among non-expert school-leaders.

Program Means: Problem-Based Instruction

Initial Acquisition of Strategic Knowledge

Our conception of how useful, strategic knowledge develops was partially in-formed by what cognitive scientists refer to as schema theory and such relatedconceptions of how knowledge is organized and stored inside one's mind asscripts (Shank and Abelson, 1977) and production systems (Sternberg and Caru-so, 1985). According to this perspective, learners commonly acquire an initialurvierstmding of (or primitive schema explaining) how to carry out some prac-tice, like solving a swampy problem, by experiencing the practice, being carriedout Ly others. This experience may take the form, for example, of a verbaldescription or the observation of modeled behavior (e.g., watching a fellowprincipal give feedback to a teacher after observing a lesson). Following Stern-berg and Caruso (1985), the learner's initial schcma or understanding is piecedtogether by identifying those elements of experienced information consideredrelevant to carrying out the practice (e.g., identifying teachers' strengths as wellds weaknesses) and putting this information together as an integrated structure orunderstanding. This new structure or understanding is then related to otherinformation already stored in memory. The resulting primitive schema or picturethen serves as a guide for the learner's initial performance (e.g., attempting,oneself, to give a teacher helpful feedback).

As a result of initial performances, the learner is provided with informationpotentially useful in refining the initial schema. This is information about discre-pancies between the actual performances and learners' schema-dependent under-standings of what ought to have happened. Such information may indicate to thelearners that: (a) their guiding schema requires further refinement and elaboration(e.g., 'I didn't describe what I saw the teacher doing in enough detain; or (b)their ability to perform does not yet match their understanding of what perform-ance entails (e.g., 'I know what needed to be said but I didn't choose the rightwords'); or (c) both.

Development of increasingly skilled performances, based on increasingly


21 6

Developing Expert Leadership for Future Schools

sophisticated cognitive schemata, depends on opportunities for repeated practiceand the quality of the feedback provided as a result of practice. A skilled coach(perhaps a more experienced colleague) already possesses a sophisticated schemato guide performance and knowledge of how best to provide feedback. As aconsequence, such a person is likely to facilitate improvement in the learner'sguiding schema and actual performance much faster than if the learner hasavailable only his own analysis of discrepancies. Burton, Brown and Fischer(1984), using the teaching of skiing as a model of instruction, conceptualizedthe conditions provided by a skilled coach in terms of Increasingly ComplexMicroworlds' (1CMs). The learner is exposed to a sequence of environments ormicroworlds in which the task demands become increasingly complex. Bymanipulating the physical setting, the equipment (where appropriate), and thetask specifications, the coach maintains a gap between the learner's initial capabi-lities and the requirements for performance that is challenging but manageable.Such a gap nourishes optimal development and refinement of guiding schemata.As Burton, Brown, and Fischer (1984, p. 148) point out: 'The goal of a sequenceof microworlds is not to remove all chances for misconceptions but to increasethe possibility that a student will learn to recognize, learn from, and correct hermistakes'.

Based on this partial conception of how useful strategic knowledge develops,it is reasonable to infer that instruction in expert school-leader problem-solvingin the swamp will be productive to the extent that it provides:

models of expert school-leader performancemultiple opportunities for practicing problem-solving in the swampa sequence of increasingly complex requirements for problem-solvingfeedback about the adequacy of performance and the sophistication ofschool-leaders' guiding cognitive schema

While schema theory contributes to an understanding of how strategicknowledge develops, it is not sufficient. Nor does such theory speak directly tothe usefulness of such knowledge. We relied on research concerned with thecontext in which learning takes place to complete our conception. This researchexplores the effects on learning of both the social context in which learning takesplace and thc purposes such learning is intended to help achieve; treatment ofthis former dimension of context will clarify our interest in developing usefulstrategic knowledge.

Most problems perceived by school-leaders to be ill-structured or swampyare defined as such as a consequence of their social rather than technical character,as we pointed out in Chapter 4. Further, most such problems have to be solvedby school-leaders in some form of collaboration with others, as discussed inChapter 6. It is particularly important, for these reasons, to understand better thesignificance of the social context for school-leader problem-solving. One defensi-ble way of thinking about the relationship between individual learning and socialcontexts accepts, as its point of dcparture, Berger and Luckman's (1966) wellknown thesis that knowledge is socially constructed, as we described in ourdiscussion of culture change in Chapter 9. This means that what counts asknowledge is socially defined, and that many of the most helpful processes used


Formal Programs for Developing Expert School Leadership in the Swamp

by individuals to acquire knowledge in practical settings involved considerablesocial interaction. As Rogoff (1984, P. 4) explains:

Central to the everyday contexts in which cognitive activity occurs -isinteraction with othcr people and use of socially provided tools andschemas for solving problems. Cognitive activity is socially defined,interpreted, and supported. People, usually in conjunction with eachother and always guided by social norms, set goals, negotiate appropri-ate means to reach the goals, and assist each other in implementing themeans and resetting the goals as activities evolve.

For purposes of the experimental program, it was most important to con-ceptualize the nature of the contributions that social interaction could make toindividual problem-solving capacity, as well as the conditions under which thosecontributions seemed most likely. One potential contribution of directed socialinteraction is the improvement of individuals' problem-solving expertise. Condi-tions required for this potential to be realized are evident, for example, inVygotsky's (1978) concept of a 'zone of proximal development'. This zone is thegap between the problem-solving capacity of the individual learner and thecapacity demonstrated by those with whom the learner interacts like colleagueswithin a group. Sensitive instruction, as Rogoff (1984) explains, at the learner'scutting edge of understanding, encourages participation at a comfortable yetchallenging level. It also provides a bridge for generalizing strategies from famil-iar to more novel situations. In this way, the problem-solving processes of thegroup are internalized by the individual. This is the same perspective used inChapter 6 to understand the conditions of group problem-solving that wouldimprove the capacity of individuals within the group.

Social interaction also has the potential for increasing the individual's capac-ity to contribute more effectively to joint problem-solving. In Mehan's (1984,p. 64) observations of committee deliberations, problem-solving was socially dis-tributed: 'The information upon which decisions are made is in the collectivememory of the group, not in any individual's memory'. In the case of theindividual learner being stimulated by the group to develop better processes forsolving problems individually, the group models performance from which moresophisticated guiding schemata can be inferred and provides socially compellingfeedback about adaptations of existing schemata. In the case of becoming a moreskillful contributor to a group's problem-solving processes, the individual isstimulated to acquire another, higher order set of schemata or understandings toguide their participation in the group.

As we pointed out also in Chapter 6 in reference to teachers, conditionslikely to foster development of such expertise among school-leaders include:

provision of group problem-solving processes likely to be more sophisti-cated than processes used by the individualopportunities for the individual to reflect on or to recover the elementsof the groups' problem-solving processesstimulation for individuals to compare their own processes to those ofthe group



Developmg Expert Leadership for Future Schools

opportunities for the group to reflect critically on the roles played byindividual members and to provide feedback to individuals about howtheir contribution could be improved

Social interaction, we argued, is a crucial feature of the context in whichproblem-solving expertise is learned (and practiced). A second crucial feature ofthe context is the nature of the problems used as vehicles for developing exper-tise. For purposes of developing practical or useful strategic knowledge amongschool-leaders, such problems must be approximately the same as the problemsthey encounter in real life and perceived as authentic by them. There are severalcompelling reasons for this to be the case. The 'situated' nature of useableknowledge is one reason for using authentic problems as vehicles for developingexpertise. As Brown, Collins and Duguid (1989) argue, for example, withrespect to what is learned, the activity in which knowledge is developed is neitherneutral nor separable from it. 'Situations' (or problems used as instructionalvehicles) along with the social context for instruction co-produce knowledge.While the instructor is not uninfluential in the process, often mediating thelearning, it is the learner, finally, who is doing the selecting. What the learnerselects, from the situation in which learning takes place, as the basis for develop-ing meaning, is unpredictable, at best. Furthermore, there will be substantialvariation among learners in their selections depending on the nature of thoseexisting cognitive structures used in the creation of meaning.

The situation (or activity) in which learning takes place, therefore, cannot beseparated from the knowledge to be learned; it is part of the knowledge that isstored in memory by the learner. The most certain way of ensuring appropriateunderstanding of what is learned, therefore, is to create situations the learner willencounter in practice. This not only aids in the initial creation of meaning, it alsohelps ensure appropriate application of knowledge subsequently: 'knowledge willbe retrieved only if the retrieval cues available at the time of access match the cuesthat were encoded with an item of knowledge' (Tulving and Thomson, cited inStemberg and Caruso, 1985, p. 149). Bransford, Franks, Vye and Sherwood (inpress) extend this idea in their analysis of conditions for knowledge transfer(discussed later).

A second reason why instructional situations and problems should beauthentic has to do with the large amounts of usually tacit and/or 'ordinary'knowledge (Lindblom and Cohen, 1979) required actually to use a generalproblem-solving strategy in practice. Authentic situations increase the probabilitythat learners will make connections with existing tacit knowledge. Artificialsituations decrease this probability. When learners take advantage of social in-teractions among peers (as when a group of five or six school-leaders try to solvean authentic problem together), the opportunities are increased for tacit know-ledge to bccome explicit and thereby examined. Further, it is thcn possible toacquire the formerly tacit knowledge of one's colleagues developed through hardexperience. This formerly tacit knowledge of one's colleagues may be easily asimportant a contribution to problem-solving expertise as the research-basedknowledge more typically the exclusive focus of formal instruction.

Finally, authentic situations are simply more motivating for learners than aresituations abstracted from the context in which knowledge is to be used (Stern-berg and Caruso, 1985). Besides being easier to relate to, familiar or real-life



Formal Programs for Developing Expert School Leadership in the Swamp

problems provide the usually enjoyable opportunity for sharing relevant andoften humorous anecdotes. Our experience suggests, as well, that the use ofauthentic problems under conditions discussed to this point, also adds to thedomain-specific knowledge and to the strategic knowledge of some participants.Receiving specific pointers about how to deal with possible problems may also bepart of what is motivating about authentic situations.

Conditions to be met by an instructional program, based on these viewsinclude:

providing instructional situations which authentically approximate thecircumstances of actual school-leader practiceencouraging the recovery, sharing, and evaluation of relevant knowledgewhich would normally remain tacit

Additional Interventions to Foster Transfer

The transfer of knowledge from the instructional setting to the real-life adminis-trative setting is assisted through some conditions already examined: in particu-lar, the use of authentic problems and settings. More can be done, however, toensure appropriate application of strategic knowledge by drawing from results ofresearch on transfer of training.

For purposes of designing the experimental program, transfer refers to:

the impact of learning a behavior [broadly defined] on the same per-formance in another context or on a different performance not simplycontaining the first in the same or a different context (Salomon andPerkins, p. 5, mimeo)

In order to design the program for greatest transfer, we were guided by anorientation to transfer, recently proposed in a series of papers by Perkins andSalomon (Perkins and Salomon, 1988, 1989, Salomon and Perkins, mimeo). Twotypes of transfer are distinguished in this formulation: 'high road' and 'low road'- each produced by different ways of thinking and stimulated by differentinstructional approaches. Because each type of transfer produces valuablealthough different outcomes, Perkins and Salomon (1988) argue that instructionalprograms ought to attempt to foster both. In general, transfer of any sort arearsto depend on how one initiates a search through one's memory for already storedclusters of knowledge that might usefully be applied to the task one faces.Transfer also depends on the extent to which such stored knowledge is linked toother potentially relevant clusters of knowledge also stored in memory.

Low-road transfer is characterized by the ready application of well-learnedknowled ge and/or skills to the same performance, in contexts very similar toones in which the knowledge and skills were originally learned, or in perform-ances only modestly different from the originally learned performances (e.g.,transferring what you know about running a staff meeting to the task of chairing

a district-curriculum committee). Such transfer occurs as a consequence of exten-sive practice in varied contexts until the application of the learned knowledge andskill becomes automatic: that is, until the presence of a situation or condition


Developing Expert Leadership _for Future Schools

elicits the application of relevant knowledge and skill, with little or no consciouscognitive mediation. Practice extends the application of what has been learnedfrom its initial performance or context to other similar performances and con-texts. Variation in the application contexts or performances (e.g., chairing manydifferent groups) during practice gradually stretches the boundaries of applica-tion, as the learner adapts to partially new circumstances in minor and largelyunconscious ways. The advantage of low-road transfer is smooth and reliableapplication of mastered performances with little expenditure of cognitive re-sources (e.g., you don't have to think much consciously). This suggests thatlow-road transfer is most appropriate for knowledge and skills subject to routineand frequent use. The 'automaticity' associated with low-road transfer, on theother hand, inhibits analytic reflection on matters such as when the application ofexisting skills and knowledge is appropriate.

Whereas low-road transfer is relatively 'mindless', high-road transfer is'mindful'. It involves applying a set of rules or principles extracted from one ormore contexts and/or performances, to other, quite novel, contexts and/or per-formances (e.g., using what you know about how to chair a meeting in order toconduct a group seminar in a graduate class). The greater the degree of abstrac-tion, the larger the range of instances such rules and principles subsume and towhich they can be applied. High-road transfer depends on 'both the decontex-tualization and representation of the decontextualized information in a new, moregeneral form subsuming other cases' (Salomon and Perkins, pp. 10-11, mimeo).This is not critically fostered by practice, although evidence suggests that limitedamounts of practice are helpful. Rather, such transfer depends on bringingconsiderable thought to bear on determining the generalizable features of theinformation. Not only must these generalizable features be determined by thelearner, but they must also be genuinely understood, evoked in the transfercontext and effectively app.ied (e.g.. beginning to understand that 'giving every-one air time in a meeting' is a way of diagnosing the existing understanding eachperson has of the issue being discussed).

Salon-ion and Perkins (mimeo) also point out that, because of the relativelymindless automaticity that comes with practice and expertise, experts may havedifficulty mindfully transferring well mastered skills and knowledge to novelcontexts. This suggests that, especially with people who have considerable exper-tise, guided abstraction from (or reflection on) well-rehearsed practices may bequite important in helping them make better use of what they already know inresponding to novel problems.

This conception of the transfer process implies a number of conditions to bemet for successful transfer of problem-solving skills. In the case of frequently androutinely used skills (e.g., information collection), most suitable for low-roadtransfer, it will be important to:

provide many opportunities for application or practice across a widevariety of problem typesprovide feedback about the adequacy of performance and opportunitiesfor further guided practice

For general principles and skills most suitable for high-road transfer, it willbe helpful to:



Formal Programs for Developing Expert School Leadership in the Swamp

provide assistance in decontextualizmg and abstracting generalizablefeatures of existing problem-solving practicesprovide direct instruction in the key components of effective problem-solving practices and coaching in the application of such components to

specific cases

Fostering Metacognition

School-leader expertise is partly a function of applying useful, strategic know-

ledge to the solution of swampy problems. To this point in the explanation of

our framework, the nature of that knowledge and how it is initially developed

has been described. We have also described how to help ensure that useful,

strategic knowledge developed in a formal instructional setting, for example, will

be applied in appropriate; real-life school-leader settings.The strategic components of administrative expertise are not limited to the

use of problem-solving strategies alone, however. Argyris' conception of double

and single-loop learning help illustrate why this is the case (e.g., Argyris, Put-

nam and Smith, 1985, Argyris, 1982). The term `single-loop learning' is applied

by Argyris to a process in which the learner responds to a problem by initiating a

chain of actions intended to resolve the problem. If that chain of actions provesless effective than desired, the learner chooses another chain of actions in a further

attempt to resolve the problem, as the problem was originally conceived. So,

while the solutions change as part of single-loop learning, the variables, gov-

erning the learner's understanding of the problem and the setting, remain the

same. In instances of double-loop learning, the governing variables are them-

selves objects of conscious scrutiny and, as a consequence, the nature of the

problem and the setting may be significantly redefined (e.g., `maybe this isn't a

discipline problem after all, maybe it's a sign of a deficiency in our program').

Individuals engaged in double-loop learning are also actively aware of their own

thinking; thcy arc involved in managing their own cognitive resources, and

monitoring and evaluating their own intellectual performance. This is inetacogni-

non (Nickerson, 1988-89). Some would also refer to this process as a form of

reflection in which experience is reconstructed in such a way as to enable one to

transform one's own practice (Grimmett, 1989). During such reflcction, the

learner engages in a 'conversation' with the setting in an effort to understand

it better and the meaning that it has for the assumptions underlying the problem-

solving activity.Conditions which seem likely to foster the development of metacognition


the provision of cues to the learner which are likely to stimulate self-

questioningthe modeling of metacognitive thinking by othersdirect instruction about the value of metacognition (since metacognitiveactivity is unlikely in the absence of a belief in its value) and the kinds ofquestions one might use as aids to self-reflection


Developing Expert Leadership for Future Schools

Instructional Strategies Used in the Program

Table 12.2 summarizes the instructional strategies used in the experimental prog-ram to address those conditions we have identified as contributing to the de-velopment of useful, strategic knowledge. As a whole, these strategies constituteour version of 'problem-based instruction' (Hallinger and McCary, in press,Boud, 1985, Bransford et al., in press). Currently used with promising results ina small number of institutions but across a wide range of professions (e.g.,medicine, management, agriculture, architecture), problem-based instructionacknowledges especially well the situated nature of cognition and dilemmasassociated with inert knowledge and lack of transfer. It does this by centeringinstruction around key problems of practice. Students acquire both domain-specific and useful, strategic knowledge in the context of working through suchproblems. The contribution of problem-based instruction can be explained interms of readier accessibility of knowledge acquired through such instruction, forapplication in practice. Such accessibility is a function of the way knowledge isorganized and stored in memory initially, around problems of practice.

The experimental program required participants to work on parts, or all, ofa total of nine problems. These problems varied in their degree of structure,although all were perceived co be relatively unstructured. In addition to degree ofstructure, two other variabks were manipulated in order to provide 'IncreasinglyComplex Microworlds'. One variable was the number of components in ourproblem-solving model (i.e., interpretation, goals, principles, etc.). Participantswere asked to address from one to all six of these components in relation to agiven problem. A second variable was the function participants were asked toperform in relation to the components. These functions included: (a) describinghow a model of problem-solving (e.g., a principal talking about how he or shesolved a problem) addressed the component(s), the simplest function; (b) evaluat-ing the model's performance; and (c) addressing the component(s) oneself, themost complex function.

Other variables manipulated through the program were the social contextfor problem-solving (individually, in pairs, in groups) and the form in whichthinking was communicated (orally, in writing or aucho-taped). Participants wererequired to evaluate the program at the end of each instructional day and asummary of their opinions was reported to them at the beginning of the nextday, along with an indication of adjustments made to the program, in response tothese opinions. This was intended, in part, to stimulate thinking about the typesof experience most helpful in stimulating thinking.

A Summary of One Day's Activities in the ExperimentalProgram

To illustrate more clearly how the conditions for optimal learning were met, weprovide a more detailed picture of a typical day in the life of our experimentalprogram. The numbers in parentheses indicate the item on Table 12.1 of whichthe commentary is an example. The agenda for each day was distributed to allparticipants when they arrived. Each session was organized around a themeusually one or two of the components in the model. On Day 2, the focus was onsetting goals and dealing with constraints.200


Table 12 1

Formal Programs Jo- r Developing Expert School Leadership in the Swamp

Development of useful, strategic knowledge: Instructional strategies andconditions

Conditions Instructional Strategies

1 provide models of expertproblem-solving

2 provide practice opportunitiesacross wide variety of problemtypes

3 sequence increasingly complextask demands

4 provide performance feedbackon individual problem-solving

5 insure participation insophisticated group problem-solving processes

6. encourage individualreflection on own and groupproblem-solving

7 provide performance feedbackon contribution of individual togroup problem-solvingprocesses

8 provide authentic instructionalsettings and problems

9 assist in recovering, sharingand evaluating tacit knowledge

10 assist individuals indecontextualizing andabstracting general features ofexisting problem-solvingpractices

11 provide direct instruction ineffective strategies andcoaching in their application

12 provide cues to stimulateself-questioning

audiotaped examples of expert administratorsdescribing the process they use to solve acase problem'live' administrators tell how they solved a realproblem they facedask individuals to write solution to ov selectedproblem; colleague critiqueask individuals to solve colleagues' problem:colleague critiquegroups of 5/6 participants solve problemstogetherproblem-solving tasks for individuals and groupsmanipulated in terms of. evaluate other'ssolution vs solve oneself; number of problem-solving components to consider, complexity ofcase problemsresponse by individual colleague, group,instructor to processes described in writing anddescribed verballycareful instruction to groups prior toengagement in problem-solving task

individual participants required to think aloudabout their own solving of problems or to writesolutions to problems. Peers and instructorsdiscussed, with individuals, their processesnot done in this program

instructional problems identified throughresearch on problems encountered by principalsand by having case problems written (or orallypresented) by administrators as they encounterthemwork with a peer or group of peers tosolve problems collaborative, discussingalternative proposals based on experience andthe thinking leading to such proposalsmost case problem-solving by individuals orgroups followed by 'debriefing' in which thecomponents of the problem-solving model wereused as the framework for discussionuse of check-lists of general strategies to beconsideredbrief presentations by instructors oncharacteristics of effective problem-solving asidentified in research; readings provided tostudents describing effective problem-solvinginstructors continuously monitored groupproblem-solving processes and intervened aswarranted to provide cues or orienting questionscheck-lists


Developing Expert Leadership for Future Schools

Table 12 1 (cont

Conditions Instructional Strategies

13. model metacoanition and 15 percent of program devoted to looking atprovide reasons for same problem from four different perspectivesmetacognition (legal, political, financial, educational)

each perspective presented by a person withspecial training or experience in the perspective(lawyer, trustee, business administrator,principal)different, lustified approaches to problem-solving and solutions evident throughobservation of different perspectivesdaily evaluation of program


Each day began with a summary of the results of the previous day's evaluation.Participants received a handout showing the frequency counts of responses to theforced-choice questions and typical responses (with frequency counts) to theopen-ended questions. They were also told how the program was adjusted inresponse to their opinions. By showing them that the evaluations were importantand had some effect, the participants were encouraged to reflect seriously on theday's events (13). The second part of the introduction reviewed the proceedingsto date and the plan for the day was unveiled.

Solution to Case Problem

On Day 1, groups had discussed how to interpret a problem that had actuallyoccurred (8) and there had been various interpretations offered. On Day 2, thearea superintendent directly involved with that problem was on hand to providethe actual interpretation and eventual solution (1). A listening guide comprised ofquestions related to the problem-solver's use of each component in our modelwas provided to focus participants' thinking (10, 12). Afterwards, the wholegroup had the chance to analyze the superintendent's solution by comparing theirown previous interpretations with those of the expert (6) and by discussing theirunderstanding of how the other components were dealt with (10).

Individual Writing: Solving a Swampy Probletn

Capitalizing on the belief that writing crystalizes one's thoughts and enhances thelikelihood that learning takes place, participants were asked to write the solutionto a swampy problem they had encountered during the preceding few weeks andwhich they had dcscribcd as part of a homework assignment. For this session,cach participant was askcd to write the solution to a colleague's swampy problemand then critique each other's solution with a main focus on the interpretation



Formal Programs .for Developing Expert School Leadership in the Swamp

component (learned on Day 1) hut also touching on all of the components (whichhad been described briefly). This activity provided an opportunity to work on anew problem (2), was moderately complex in terms of the number of compo-nents addressed and the difficulty of the task (3), and provided performancefeedback on individual problem-solving (4). The discussions between colleaguesencouraged reflection about problem-solving in general (6), provided an oppor-tunity to share tacit knowledge (9), and also provided direct instruction aboutspecific strategies that worked well or failed miserably (11). Additionally, theproblems dealt with were obviously authentic (8). This activity also met acriterion for optimal learning not considered on Table 12.1: the participantsfound these discussions enormously enjoyable.

Setting Goals

This section represents our typical method for presenting new information. Ageneral description of the meaning and importance of the goals' component wasdelivered orally by an instructor; the group also received a one-page handoutsummarizing the talk (11). A written case problem which was particularlyappropriate for illustrating the importance of identifying goals was distributed (2)and possible solutions were discussed in small groups according to instructions atthe bottom of the one-page case description (5). Those instructions pertainedespecially to the goals that a solver would set but asked for considcration of allother components (3), as well. Following this, a check-list detailing the kinds ofthinking in experts engage which regarding goals was distributed (12). Smallgroups reanalyzed the 'implementation problem' in light of these specific cues.These reactions were then shared with the whole group (10). A similar procedurewas used for the 'constraints component' using another problem entitled 'theMarks Scenario' (2). Two errors commonly committed by non-expert problem-solvers were described (11) and the groups' solutions to the 'Marks Scenario'problem were examined for evidence of them (10). Finally, an expert problem-solver and educational leader described how he would have solved this problem(1) with the group engaged in focused listening, using the guide described earlier(10, 12). The entire group was invited to ask specific, content questions (9) anddiscuss the strengths and weaknesses of the solution process (10). To end the day,homework related to the next day was assigned, a relevant article was distributed(11), and participants evaluated the day's events (13).

In one day, all of the conditions we have listed for stimulating learning hadbeen arranged for at least once (and most more than once). What is missing fromthis narrative is the importance of keeping the group keenly motivated. Carefulpacing of activities to keep interest high and a wealth of illustrative cartoons wereessential elements of the program.


Evidence collected to evaluate the program described in this chapter (e.g., Leith-wood and Steinbach, in press) supports the claim that even quite experienced


Developing Expert Leadership for Future Schools

school-leaders are abk rn .nrrpice rhpir Pvnprtice+ is-. 1 1


Chapter 13

Performance Appraisal andSelection of School-Leaders:Performance Appraisal for Growth

Formal programs, as described in the two previous chapters, are obvious andquite direct strategies for fostering school-leader development. School-leaderselection and appraisal policies and procedures, while possibly less obvious de-velopment strategies, are also quite direct.

In the case of selection procedures, it is the school's rather than the indi-vidual's leadership capacity which is being developed most directly through thechoice of the greatest actual or potential expertise. Especially during periods,such as the 1990s, of unusually rapid turnover in school-leaders due to retirement(Leithwood and Begley, 1985, Musella and Lawton, 1986, Peterson, Marshalland Grier, 1987) these choices are of enormous consequence to the quality ofleadership which will be available to schools as they enter the twenty-firstcentury.

Appraisal policies and procedures, as they are used 'summitively' to assistwith dismissal, promotion, and transfer decisions, influence the school's leader-ship capacity in approximately the same way as do selection practices. Whenappraisal policies are used 'formatively,' they are capable of fostering the growthof individual school-leaders in precisely the manner we outlined for teacher-appraisal policies in Chapter 8.

This chapter and Chapter 14 apply a single framework to issues involved inusing appraisal and selection processes to foster the development of leaders forfuture schools. In the case of each sct of processes, we review research describingcurrent practices and then explore ways of increasing the contribution of suchpractices to leader development. Such improvement is often considered to be amatter of improving the quality of the information available to appraisers andselectors. And, as we shall see, there is much to be done to bring typical appraisaland selection processes 'up to speed' with regard to such quality. A majoremphasis throughout these two chapters, however, will be on making effectiveuse of the information that is available. Compelling evidence suggests thatinformation generated during appraisal and selection processes is often not usedor is misused (see, for example, Hickcox, Lawton, Leithwood and Musella,1988). Under such conditions, the contribution of appraisal and selection pro-cesses to school-leader development is likely to remain severely blunted. Greaterattention to use, we argue, is a promising avenue for improving appraisal andselection processes.


Developing Expert Leadership for Future Schools

This chapter is concerned exclusively with performance appraisal. We beginby examining thoroughly current practice, by reviewing recent literature on thetopic, and then report on some of our own research concerned with identifyingeffective practices. In Chapter 14 we examine current selection processes and thenconsider a number of issues involved in the measurement of school-leader per-formance. These issues, while they may not appea: to all readers due to theirinherently technical nature, have significant implications for both school-leaderappraisal and selection. A summary of guidelines for improving both sets ofprocesses is left to the end of Chapter 14.

Knowledge about current school-leader selection and appraisal practices isquite limited. Our attempt to find and review relevant knowlelge included theuse of computerized searches (ERIC), and scanning recent issues of relevantperiodicals and bibliographic follow-up. Only a few comprehensive, descriptivesurveys of selection and appraisal practice were found. Most studies focused onthe utility of participant forms of selection and/or appraisal (e.g., performance-assessment centers) rather than engaging in more comprehensive reviews ofpractice. Furthermore, most of what has been written about sekction andappraisal practices might best be termed 'opinion' literature. It may be that thebulk of this writing is based on valuable and illuminating practical experience butthe reader rarely has a way of making this determination. Hence, mention ismade of this literature only sparingly and when judged to overlap significantlywith research findings.

Our reviews of both the appraisal and sel:ction literature are organizedaround a conceptual framework used to guide the research of Lawton, Hickcox,Leithwood and Musella (198(). The framework includes five process-orientedcomponents:

Preparation: planning, purposes, criteria, and standards for selectionand appraisal. Essentially activities that occur prior to the evaluation thatdetermine its form and substanceData Collection: sources and types of information, methods of collec-tion, who collects data and how much time is spent in doing soReporting and Follow-up: the nature of the report, the form it takes,its destination, the development of plans of action and other conse-quences of the data-collection processEvolution of Policy: processes by which policies are developed, im-plementation activities, and whether policies are reviewed and revised ona periodic basisImpact: degree of compliance with policy, commitment to implemen-tation and administration, and the nature and degree of impact withinthe system, including esp...ially 'use'.

The Current Status of School-Leader Appraisal

Much has been written about the appraisal of school-leaders but, as with theliterature on selection, research in the area is not well developed. Based on acomprehensive literature review, Ginsberg and Berry (199(1) suggested that muchof what is written on school-leader evaluation concerning models, procedures.

2062 ,2

Performance Appraisal for Growthand instruments is description of local practice and not explicitly supported byempirical research. Results of the few significant and relevant studies which wereavailable support Ginsberg and Berry's (1990) suggestion that school-leader eva-luation practices are in thc 'stone age'.

The description of current assessment practices presented in this sectionrelies most heavily on the study by Lawton et al. (1986). This study included ananalysis of existing policies in Ontario, a survey of thirty school systems withdata from 4092 teachers, 879 principals, 114 supervisory officers, and twenty-sixchief education officers (CEOs). Also included were a series of case studies basedon interview data witb local administrators and teachers in thirteen school dis-tricts. Also reflected in our description of current appraisal practices are theresults of a comprehensive (but still descriptive) study reported by Duke andStiggins (1985). They surveyed one district administrator with responsibility forschool-leader evaluation and two principals (one elementary and one secondary)from each of thirty school districts. Other studies reviewed for this sectioninclude Ginsberg and Berry (1990), Berry and Ginsberg (1988), Bolam (1990),Duhamel, Cyze, Lamacraft and Rutherford (1981), Harrison and Peterson (1986),fancy (1988), Murphy, Peterson and Hallinger (1986), Peters and Bagenstos(1988). and Redfern (1986). Comments from advocates of various approaches toschool-leader evaluation (Ingle, 1975, Rentsch, 1983, Valentine, 1987) are inter-jected where appropriate.


Preparation for school-leader appraisal includes planning, perceived purpose, andcriteria-setting activities. Ginsberg and Berry (1990) reported that not manyschool systems had formal policies. This was corroborated by Lawton et al.(1986) although these authors believed that the trend was quickly changing.Documentation for school-leader appraisal policis was found to be more preva-lent than for other roles, and most districts either supplied formal policies orindividuals responding to the survey were able to describe policies in writtenform.Considerable variation existed in the procedures for informing school-leaders about the district's appraisal process (Duke and Stiggins, 1985). Theseinclude relevant communication mechanisms, handbooks, written instructions,and verbal explanations. Lawton et al. (1986) found that over half of the princip-als they surveyed received personal notification that a formal appraisal waspending and that pre-conferences between principals and appraiscrs were quitecommon (over 70 percent) Pre-conferences were used primarily as methods forsetting objectives for the impending observation period (usually throughout thecourse of the year).Ginsberg and Berry (1990) reported that the most common purpose ofschool-leader appraisal was performance improvement (formative). This purposealso appeared to be most common among those engaged in headmaster-appraisalpilot projects currently underway in the UK (Bolam, 1990). Ingle (1975) sug-gested that although improvement and decision-oriented (sunimative) purposesfor administrative evaluation can be separated conceptually, this cannot be donein practice. Indeed, research on currcnt practice showed that there may be

2 0 207

Developing Expert Leadership for Future Schools

considerable confusion about these fundamental functions of school-leader

appraisal. But not all available data support the claim that improvement was the

acknowledged purpose of performance appraisal. When asked about actual and

preferred purposes for appraisal, principals were strongly inclined to indicate that

although performance improvement (variously referred to, as supervision for

growth, formative evaluation, developmental evaluation, etc.) was highly

valued, it did not appear to be the most commonly acknowledged function in

practice (Duke and Stiggins, 1985, Lawton et aL, 1986). Lawton et al. found that

'comply with policy' was the most frequently identified actual purpose. Some

would suggest that 'comply with policy' is not a purpose at all and interpret this

finding as revealing principals' lack of true understanding of the intended pur-

poses of the policy and/or cynical attitudes toward organizational control and

accountability (Hickcox, 1990). They found, and others as well (e.g., Duke and

Stiggins, 1985, Peters and Bagenstos, 1988), that considerable weight was given

to summative purposes such as tenure, transfer, promotion, certification, and

retention remediation. Harrison and Peterson (1986) reported that principals' and

superintendents' views were somewhat discrepant about principal appraisal.

Whereas superintendents were quite positive and more consistent in their percep-

tions, principals were less clear about the intentions of the process and in less

agreement about them. Finally, in many states in the US merit-pay plans were

being implemented (Duke and Stiggins, 1985, Berry and Ginsberg, 1988). Such

evaluation plans constituted another example of decision-oriented or sumrnative

purposes for appraisal.Criteria for evaluating school-leaders were extremely varied both across and

probably within school systems. Murphy et al. (1986) found that in effective

school districts, school-leaders were held accountable for school objectives, an

observation they believed may be attributable to an instructional management

orientation of supervisors. Other authors observed that 'management by object-

ives' models were the most popular forms of principal appraisal (e.g., Ginsberg

and Berry, 1990, Peters and Bagenstos, 1988, Redfern, 1986). Ginsberg and

Berry (199(i) indicated that, in addition to outcome-oriented evaluations, a popu-

lar focus for principal appraisal appeared to be observable behaviors or processes

in which principals engage. The development of behavioral rating inventories for

use in appraisal activities is consistent with this orientation to evaluation Varia-

tion in criteria and lack of defined performance standards were common in

current practice:

no consensus among districts exists concerning the precise nature of

instructional leadership or school management. Forty-two percent of the

respondents ... cited inadequate performance standards as a shortcom-

ing of the existing principal evaluation system. (Duke and Stiggins,

1985, p. 92)

The findings of Lawton et al. (1986) were consistent with these results since

appraisal criteria were shown to be vague, ambiguous, and not necessarily linked

to any existing conception of expert practice. On the other hand, they did

observe that as compared with other roles, the criteria used for principal appraisal

were more broad and focused on activities both in and outside the school. Also,

they reported a tendency tbr standards or performance expectations to be de-



Performance Appraisal fur Growth

veloped collectively with appraisers, a practice commonly associated with moreimprovement-oriented appraisal. Other studies reported criteria and standard-setting activities to be unclear, inconsistent and troublesome (Ginsberg andBerry, 1990, Harrison and Peterson, 1986). As several authors have suggested, inaddition to outcome and process, other criteria for evaluating principals mightfall into a category called 'presage' which would include interpersonal skills (e.g.,appearance, personal characteristics, and traits). The use of such criteria, it wouldappear, is rapidly diminishing (Duhamel et al., 1981, Ginsberg and Berry, 1990,Rentsch, 1983). Some advocated approaches to principal evaluation underscorethe need to link process variables to outcomes (e.g., Valentine, 1987). It is notclear, however, to what extent this has been accomplished in practice.

Data Collection

Typically, school-leaders are formally appraised, on an annual basis, by theirimmediate supervisor (Lawton et al., 1986). Some variation in this tendencyexists and informal ongoing supervision by immediate supervisors is becomingmuch more prevalent (Duke and Stiggins, 1985): it is also characteristic ofprocesses in effective school districts (Murphy et al., 1986). School-leaders play acentral role in providing data for their own appraisal, according to Lawton et al.(1986). Data from supervisory officers was found to be a distant second source ofinformation about principals. Relatively infrequently, teachers, parents, and stu-dents were sources of data. This finding corroborates observations by Ginsbergand Berry (1990) and Redfern (1986) who noted that rarely do appraisal processesinclude multiple data sources as a standard feature for data collection. In caseswhere this occurs it is likely to be optional (Redfern, 1986). Bolam (1990)reported that peers are used in thc newly developing, national appraisal processesin the UK but that confusion exists concerning the role to be played by variousindividuals. Keck and Hampton (1987) advocate a portfolio-based peer-assessment model where multiple assessors help to ensure adequate reliability andvalidity.

Janey (1988) advocated the use of parents as a source of data for principalappraisal but showed that current school-leaders are not comfortable with thisapproach, presumably because they would question the validity of parentobservations. Harrison and Peterson (1986) reported discrepancies in views heldby principals and supervisors concerning sources of data for current appraisalpractices. While supervisors made the claim that they used multiple sources ofdata and, especially, information derived from the principals themselves, princip-als suggested that observations from the community were the appraiser's primarysource of information. Indeed Murphy et al. (1986) confirmed that such data, inaddition to student-achievement data, were used to evaluate principals in effectiveschool districts. However, the authors acknowledged that these sources wereused as 'perception-checking activities' as opposcd to assistance in making judg-ments. For thc latter, superintendents tended to rely on their own observations ofthe principal and available quantitative data. For this reason, supervisors in theeffective districts in the study necessarily made frequent visits to the schools.

The most frequtflitly identified types of data were school-evaluation reportsprepared by principals followed by notes taken by supervisors. and documents

232 209

Developing Expert Leadership for Future Schools

such as school-communit), newsletters, according to Lawton et a/. (1986). Severalother types of data including behavior check-lists, written self-evaluations, schoolbudgets, timetables, and handbooks were also examined but much less frequent-ly. Duke and Stiggins (1985) reported that planned and unplanned observationsby superintendents were the most frequently identified types of evidence used forappraisal but that these were less likely to be used than informal input fromteachers, other school personnel, and especially parents. Principals were mostlikely to be observed in meetings with community organizations and facultymeetings (see also Murphy et al., 1986). Bolam (1990) reported that head-teacherswere generally observed for both management and teaching performance, sinceteaching represents a significant component of their responsibilities. Bolam alsoindicated that interviews with colleagues appeared to be a popular method ofcollecting data in the early stages of the national-assessment program theystudied. As Ginsberg and Berry (1990) noted, there are many instruments andcheck-lists in existence and probably in use but not an abundance of informationconcerning their use. School-leaders often believed that instruments such ascheck-lists and behavior inventories are used too much in practice; in contrast,district administrators preferred to see them used more (Lawton et cd., 1986).

In the Lawton et al. (1986) study, information was generally collected andevaluated by a single district administrator, usually the principal's area super-intendent'. In addition, information was usually collected in one day or less.Although there was some variability in this trend, over 50 percent of the samplereported this to be the case. In a few cases, a team of supervisors or appraisers (insome instances CEOs were part of the team) were responsible for the evaluation:data were usually collected over about a week in a very intensive effort by theteam.

Reporting and Follow-up

Duke and Stiggins (1985) and Lawton et al. (1986) both reported heavy relianceon the use of post-conferences between the appraiser and principal following theformal evaluation. Typically, a written report was prepared and either forwardedto the principal prior to the conference or brought to the meeting by thesupervisory officer. This written report, which generally did not exceed one ortwo pages (Lawton et al., 1986) formed the basis for discussion during themeeting. It gave the principal an opportunity to counter or even refine observa-tions by the appraiser. The usual length of the post-conference was approximate-ly one hour, on average, and in some cases more than one district administratorwas present.

Lawton et al. (1986) reported that principals' perceptions of post-conferenceswere very favorable. For the most part, principals found their appraisers to befair, sincere, and to take the process seriously. The post-conferences were foundto be positive and non-threatening experiences. On the other hand, principals didprofess to take the process seriously. These findings leave one with the sense thatthe post-conference is basically a 'summing up' of the supervisors' observationsand that no surprises were typically apparent. Ingle (1975) noted that trustbetween appraiser and school-leader is an important feature of reporting and

2109 3 3

Performance Appraisal for Growth

follow-up but suggested that a certain amount of pressure is necessary if per-formance improvement is expected to occur.

Bolam (1990) noted that statements about head-teachers were being preparedand sent to CEOs following the appraisal and that head-teachers were feelingsomewhat anxious about this. Finally. Ginsberg and Berry (1990) indicated thatthe quality of communication between appraiser and principal was typically notvery high.

Evolution of Policy

Lawton et al. (1986) observed that principal-evaluation policies were generallydeveloped by district administrators with some input from principals. Suchpolicies were typically less specific than teacher-evaluation policies but morespecific than policies developed for superordinate roles. Systematic revision oreven cyclical review of policies appears to be lacking in most districts accordingto the authors. There was neither much evidence of systematic monitoring ofhow well such policies were implemented nor much effort to 'fine tune' policiesonce developed. Some policies took years to develop where others had beenconstructed in a hurried fashion by adapting policies developed in other districts.The authors concluded that impact was likely to be higher where policies wereconstructed from the ground up or where significant local energy was invested intheir design and implementation.

Other than the Lawton et al. study, we were unable to locate data on thetvolution of principal-appraisal policies.

One of the most interesting findings of the Lawton et al. (19M) study was that,although 'satisfaction' with the appraisal process was comparatively high, impactwas low. -I he extent to which performance improvement was believed to be aconsequence of the appraisal was minimal, as reported by both principals anddistrict administrators. This is not surprising since in only about 25 percent of thecases were individual development plans developed by, or for, the school-leadersas a consequence of the process. Such lack of attention to follow-up helps explainthe results mentioned earlier, that appraisal for performance improvement wasperceived by many as the desirable but not thy actual purpose for doing theappraisal in the first place.

In addition, m very few cases were administrative actions (e.g., promotion,dismissal) a consequence of the process. Less than 2 percent of principals wereplaced on review as a result of the process and about as many received some sortof recognition or award as a consequence. These data largely corroborate those ofDuke and Suggins (1985). Perhaps the most frequently observed outcomes of theappraisal processes they studied were individual and private acknowledgementsof satisfactory or more than satisfactory performance. Even though principalstend to he very satisfied with the processes used in their appraisal, the impact of


Developing Expert Leadership for Future Schools

these processes is almost non existent. No news, it seems, is better than badnews. And good news seems not to be expected.

Finally, Berry and Ginsberg (1988) reported that principals believed thatevaluation systems structured around the merit-pay concept have neither thesensitivity to detect substandard performance nor the integrity to identify onlydeserving school-leaders as meritorious.


To this point, we have reviewed the small amount of empirical research that wasavailable, describing current school-leaders' appraisal practices. This evidencesuggests that until recently such practices have been poorly developeo and havenot been the object of systematic attention by a large proportion of schooldistricts. This portrait of neglect may well be changing rapidly, as we speak,however. The increasing importance attached to school leadership and its de-velopment seems likely to be a central force in stimulating such change. Never-theless, even the best of current practice has some distance to go if it is tocontribute to school-leader development as it has the potential to. Much greatereffort will need to be devoted to reshaping these practices so that they reinforce adefensible image of future school leadership and so that they become morepowerful stimulants to school-leader growth. We now turn to issues involved inbringing about such changes.

Ped.ormance Appraisal for Growth

Although data were exploratory and correlational in nature, 'a pattern of activ-ities and procedures emerges from the data that indicates generic techniques andstrategies do exist which are more likely to succeed than other widely usedpractices' (p. 2). Summarized here are the results that specifically apply to the

evaluation of principals.In this study, data were aggregated to the district level (number of

districts = 30) and two sets of impact measures were constructed based on datafrom various sources. The first set of impact measures concerned 'interveningeffects' including satisfaction with report form, fairness of appraiser's judgment,and fairness of procedures. Intervening effects also included: skillfulness ofappraiser, how seriously the principal took the appraisal, and how seriously theappraiser took the appraisal. The second set of impact measures concerned the'final effects' of school-leader appraisal and involved data from teachers andsuperintendents in addition to school-leaders. Included as 'final effects' were theeffectiveness of the principal appraisal reported by CEOs, and the extent ofimprovement in the principal's performance, as reported by the principal,teachers, and CEOs. Lawton et al. used the framework described in Figure 13.1

to summarize the findings.

PreparationThe existence and length of pre-conferences with principals were positivelycorrelated with a number of intervening and 'final effect' variables. Other vari-ables such as method of notification and use of 'objective-based' evaluation didnot yield consistent results. A variable that correlated positively with all interven-ing effect variables was whether purposes for the appraisal were clearly given toprincipals. Interestingly, although correlated with satisfaction measures, this vari-

able did not correlate with 'final effect' variables.Criteria such as administrative performance, school/community relations

and personnel management were not found to yield interpretable patterns ofrelationship. However, if principals' contributions at the district level were in-cluded as criteria, there appeared to be a tendency for the appraiser's judgment tobe perceived as fair and for the superintendent to perceive the performance-appraisal system as being effective.

Data collectionSeveral variables in the categories: types of information used, who collectedinformation, and time spent collecting it, were tested against the impact mea-

sures. In districts where teachers were called upon to provide information for the

principal's appraisal, there were positive relationships with both satisfaction and

outcome variables. General and specific note-taking appeared to be a data-

collection practice that was accepted as satisfactory and correlated with principals'perceptions of performance improvement due to the appraisal. Findings weremixed concerning the use of student disciplinary records as a basis for appraisal.Principals appeared to take the process more seriously when such data were used

but there was a negative correlation with performance improvement. Other typesof data such as check-lists, self-evaluation questionnaires, school handbooks, andgoal packages did not yield interpretable patterns of correlations.

The use of a team of appraisers appeared to have a strong relationship withthe seriousness with which the principal took the process, as well as with thc



0 3"144 4

Figure 13 1 Conceptual framework of evaluation utilization







Decision orPolicy setting


(Politicalclimate uompeting


User commitment and/orreceptiveness to evaluation

User's personal\[characteristics,

Evaluation Utilization








Peiformance Appraisal for Growth

superintendent's perceptions about performance improvement. Greater teacherinvolvement, beyond the level of providing information, also correlated positive-ly with perceived fairness of the process and the skill with which it was carried

out. Where more time was spent collecting data, both appraisers and appraisal-subjects appeared to have taken the process more seriously. A similar, yet

stronger relationship was found where post-conferences were held. This wasrelated to principal satisfaction and perceived improvement.

Reporting and follow-upClearly, provision for holding a post-conference at the conclusion of the processturned out to be a worthwhile activity, especially in relation to its effectiveness(performance improvement, CEOs perception). As noted above, longer confer-ences seemed to be more worthwhile. There was mild evidence associated witheffects concerning the form in which communication of results occurred. On theone hand, semi-stnrctured and unstructured statements about principals' per-formance appeared to have a positive relationship to several outcomes, but theuse of rating scales was found to correlate negatively with principals' propensityto take the process seriously. Finally, although the existence of an appeal processdid not correlate significantly with measures of impact, the development of aplan for performance improvement correlated with several measures of bothintervening and performance-improvement effects.

Evolution of policyLawton et al. reported that steps taken to develop and implement policy had amodest relationship to its effectiveness. Specifically, the fidelity of implementa-tion documentation, training, and commitment of resources appeared to beimportant.

The authors urged caution in interpreting the findings of their study for tworeasons. First, the sample size of thirty school systems may be representative ofsystems in Ontario but it poses restrictions on the statistical analyses employed.Second, the analyses are only correlational and are therefore only suggestive ofpossible causal relationships. Nonetheless, the study was very comprehensive andallowed for a test against impact of a wide range of features and activities inschool-leader performance appraisal.

Perhaps the significance for performance-appraisal practice of findings fromthis study might best be summed up in the words of one of its authors:

I've developed a strong sense that what is imbedded in these findings forboth teachers and principals is that what they look for in performanceappraisal is contact with someone who has an interest in their work. It isthc conferencing aspects, pre and post, that are crucial, in the eyes ofthese respondents. If this is accurate, it has a whole lot to say about howappraisal procedures should be carried out and about the supervisoryroles. (Hickcox, 1990)

To be sure, as we shall see below, the role played by the dynamics of therelationship between supervisor and principal is crucial. Credibility, trust, open-ness, and willingness to share information appear to be integral components ofespecially growth-oriented appraisals.


Developing Expert Leadership for Future Schools

Study Two: Cousins (1988)

Cousins' (1988) study inquired about factors that influence the use of performance-appraisal data for developmental purposes by school-leaders. Do principals useperformance-appraisal data concerning their own performance? If so, underwhat conditions? The research consisted of three substudies conducted in sequence,each probing more deeply into the utilization process.

Substudy A: Current knowledge about evaluation utilizationIn Substudy A (also reported by Cousins and Lcithwood, 1986) we sought todevelop a justifiable conception of how evaluative information (of any sort) isused and what determines the likelihood that it will be used. To do this, wereviewed and synthesized a large body of original research (eighty studies) underthe general topic of 'evaluation utilization'. Most of these studies were about theuse of information resulting from program, student, course, and curriculumevaluation; very few studies dealt with the use of personnel-evaluation data.Studies were identified through computer searches and bibliographic follow-up.Each study was systematically analyzed and information such as measures of use,variables affecting use, sample, instruments, theoretical framework, and resultswas summarized. As a product of this analysis, we constructed a framework(Figure 13.1) for understanding knowledge use that could be applied to the morespecific domain of performance appraisal.

In this study 'utilization' meant both use for decision-making (instrumentaluse) and use for less direct, educative purposes (conceptual use). In ou- definitionof use, we also included: just thinking about the information (cognitive process-ing) a minimum condition for use. Two clusters of faclors were identified asmajor influences on use: characteristics of how the evaluation is implemented andcharacteristics of the decision or policy setting. Factors associated with each ofthese categories include thc following:


Characteristics of the Evaluation Implementation

Evaluation quality:Characteristics of the evaluation process includ; lg its sophistication,rigor, availability of follow-up, and the like. An evaluation that attemptsto link program components to program outcomes, for example, isconsidered to be more sophisticated than one that merely describesoutcomes.Credibility:Credibility of the evaluator and/or the evaluation process defined byobjectivity, believability, appropriateness of criteria, and so on. A well-seasoned evaluator with a proven track record is attributed higher levelsof credibility than a novice, for example.Relevance:The extent to which the evaluation is relevant (usually meaning practical)to the needs of the audience. Do the purposes of the evaluation meet theexplicit and implicit needs of the audience for whom the evaluation isconducted? Do evaluators working within the organization tcnd to pro-duce evaluations that are more relevant?

2 0

Performance Appraisal for Growth

Communication quality:The clarity, style, readability, flair, and the like, with which the evalua-tion information is communicated to the intended audience. Com-munication can be oral, written, visual, and so forth and evaluatoradvocacy of findings and recommendations, and follow-up activities areactivities that improve communication quality.Findings:The nature of the actual information (data) being disseminated: con-sistency with existing knowledge, congruence with expectations, posi-tiveness, scope, value for decision-making, and the like.Timeliness:The extent to which evaluation information is disseminated in a timelyfashion. Sometimes this means whether information is delivered in timefor impending decisions or on an ongoing basis, depending upon


Characteristics of the Decision or Policy Setting

Information needs:The type of information sought, number of evaluation audiences withdiffering information needs, time pressure, and perceived need for eva-luation. To what extent are explicit and implicit needs for evaluationinformation shared among different audiences?Decision characteristics:Characteristics of decisions associated with the evaluation problem in-cluding decision-impact area, type of decision, program novelty, andsignificance of the decision, among other examples. Decisions regardingpolitically sensitive or controversial issues are of relatively high signi-ficance.Political climate:Characteristics associated with political climate such as political orienta-tion of commissioners of the evaluation, dependence of decision makerson external sponsors, inter and intraorganizational rivalries, budgetfights, and power struggles. Is it politically prudent for decision makersto decide in a manner that is consistent with the evaluation results?Competing information:Information from sources beyond the evaluation, relevant to the researchproblem and competing with evaluation data to inform decisions. Per-sonal experience, informal observations made by decision makers, and'working knowledge' are examples.User's personal characteristics:Decision makers' organizational role, information-processing style,organizational experience, and social characteristics, among other vari-ables, fall into this category. Decision makers who carefully plan for thefuture and take preventative actions are distinguished from 'crisis mana-gers' who operate on a more 'reactive' basis.User's commitment/receptiveness to evaluation:The extent to which decision makers are open-minded about decisionsand about the evaluation findings. Arc the decision makers dogmatic



Developing Expert Leadership for Future Schools

about the decision? Are they predisposed to attitudes about the utility ofevaluation?

Substudy B: Multiple-case studyIn Substudy B, the framework was applied in an actual performance-appraisalsetting. An independent judge (district adminstrator) selected principals in fourelementary schools for the study. Two were selected because they were thoughtto be 'high users' of their own appraisal data, the remaining two were thought tobe low users'. At each of the four schools we conducted interviews about theprincipal's actions. We talked with principals, vice-principals, superintendents,teachers, and parents with a reasonable knowledge of the principal's daily activ-ities. The interviews and subsequent analyses of them were directly guided by theframework developed in the preceding study. We used analytic procedures out-lined by Miles and Huberman (1984). Also, questions were asked about theperformance-appraisal policy used to guide each principal's appraisal, in additionto its implementation and its usefulness to the principal (impact). We thenattempted to identify factors that enhanced and/or inhibiied impact.

The use of appraisal data by principals in this study was fouad to be, at best,only moderate. Very little conceptual development regarding performance im-provement was observed, even though the data appeared to be taken seriously(e.g., processed) by principals. Learning that did occur was typically a reinforce-ment of the principals' existing knowledge as opposed to them learning some-thing new. Speaking about a year-end summation of a specific appraisal, anappraiser said,

I think the appraisal process was a confirmation of what [the principal]already thought. All we did was confirm each other's impressions in thatexperience. There were really no surprises.

Similarly, decisions based on appraisal data usually confirmed impressions aboutexisting school-level programs and directions. A principal decided, for example,to carry on with staff-development initiatives focused on standardizing studentevaluation within and across grades. In some instances, there were decisions tomove in new directions. One principal decided to pursue Ontario supervisory-officer certification partially because of constant encouragement from herappraiser.

In sum, although principals tended to take the process seriously, they did notgenerally learn anything from, nor base decisions upon, appraisal informationconcerning their own performance. Why was this the case? Why didn't principalsuse appraisal data? As part of the answer, we found that principals' motivation,experience, and attitude toward appraisal limited learning attributable to it.According to one seasoned principal:

I don't remember looking forward to thc appraisal process as a learningexperience for me. I looked forward knowing it was going to be thereand very confident of the way it would be scored.

In contrast, a superintendent spoke of a somewhat less experienced and moremotivated principal:



Performance Appraisal for Growth

[The principal's] attitude is much more of 'What can I learn from theevaluation? Let's not even do the evaluation unless it's going to help meto do my job'.

Other reasons had to do with whether an open, honest, and trusting relationshipexisted between principal and appraiser. Principals who knew they would not bereprimanded for making mistakes were relatively open about discussing prob-lems and ideas with supervisors. They felt confident that the purpose of theappraisal process was to help them to improve and were not concerned aboutthings they had disclosed 'coming back' on them. Finally, the appraisers' persist-ence and the positive manner in which they communicated their findings hadrelatively small but positive effects on use.

Perhaps one of the more interesting findings of this study was that weidentified variables that influenced use that could not readily be accommodatedby out- framework. These variables were associated with the relationship betweenappraiser and appraisal-subject, and had to do with trust and willingness todisclose performance-related information. It is likely that these variables emergedbecause the user of the appraisal information is also the focus for evaluation in a'performance appraisal for improvement' context.

Substudy C: Statistical reanalysisFinally, we wanted to assess the confidence one could have in the results ofSubstudy B by using independent data. In this study, we reanalyzed a set ofquestionnaire responses from almost 900 principals in the Lawton et al. (1986)

study. As we noted earlier, one of the more dramatic findings of the survey wasthat although principals believed their appraisal policies to be fair and werereasonably satisfied with them, they were perceived to have little or no impact onperformance.

From the survey data, we examined some of the findings from our multiple-case study by investigating relationships between measures of use and variablesbelieved to affect them. We used LISREL ( Joreskog and Sorbom, 1981, Version6) to conduct path analyses on the survey data. Our knowledge-utilizationframework and findings from Substudy B guided the statistical analyses. Themarkedly low degree to which appraisal data were used by principals for thepurposes of performance improvement was confirmed by Substudy C. Suchinformation was taken seriously but nothing much happened beyond that. Thepositive manner in which information was communicated to principals, andprincipals' positive attitudes or receptiveness to the process were found to lead togreater impact. Perhaps more to the point, where these variables were negative,use of the information was minimal or non-existent.

It is evident from these studies that although the context in which appraisaloccurs is important, there are features associated with the implementation ofperformance-appraisal policies that can determine whether or not they contributeto the development of school-leaders. This is encouraging news for those re-sponsible for implementing such policies. In addition, the studies underscore the

need to work hard at developing positive interpersonal dynamics between super-visor and principal. Improvement along these lines seems likely to predict moreeffective growth-oriented appraisal practice.


Developing Expert Leadership for Future Schools


The sequence of studies reported in the last half of this chapter offer promisinginsights regarding the improvement of school-leader appraisal practices. We leavea summary of these insights to the last part of the next chapter.


Chapter 14

Performance Appraisal and Selectionof School-Leaders: Selection Processesand Measurement Issues

This chapter continues our analysis of how performance appraisal and selectionprocesses can be carried out so as to contribute significantly to the developmentof futur.' school-leaders. The first section of the chapter reviews research litera-ture about selection processes using, as a framework, the same categories ofactivities which structured our analysts of appraisal processes: preparation, datacollection, reporting and follow-up, evolution of policy, and impact. The secondand third sections of the chapter explore, respectively, issues in measuringschool-leader performance and guidelines for improving school-leader seiection

and appraisal.

The Current Status of School-Leader Selection Processes

Given the loss of leadership experienced by a school district as a consequence of a'wrong' selection decision and the problems often associated with terminating ordemoting individuals from a post, one would expect research on the topic to beextensive. However, our search located relatively little. Schmitt and Schechtman(1990) encountered thc same vacuum in their evidently careful and thoroughsearch: they were able to locate only six studies that reported predictive validitydata; most of the empirical work they found was descriptive in nature. Two suchstudies (Baltzell and Dentler, 1983, Bryant, Lawlis, Nicholson and Maher, 1978)were included in our review. In a province-wide survey of practice in Ontario,Musella and Lawton (1986) analyzed both existing policies and documents con-cerned with the selection of school-leaders (as well as other roles), and surveydata about policy implementation and impact. Archival data were obtainedfrom ninety-nine of 136 districts in the Province, and 1353 department heads,

vice-principals, principals, and district administrators submitted completedquestionnaires.

Two other descriptive studies originated from the United Kingdom. Mor-gan, Hall and Mackay (1983) observed actual selection processes and conducted acomprehensive survey about secondary head-teacher selection practices in theUK. Elkins (1987) focused more directly on what appeared to be occurring atthe level of the local education authority (LEA). In a study by Parkay and Currie(1989) desired and actual sources of support for secondary-school principals



Developing Expert Leadership for Future Schools

during selection and entry were examined. Several other studies dealt specificallywith the use of assessment-center technology in the selection and development ofschool administrators (Allison and Allison, 1989, 1990, Allison 1989, Bryant,1990, Gomez and Stephenson, 1987, Milestein and Fiedler, 1989; Nagy andAllison, 1988, Schmitt, Noe, Meritt, Fitzgerald and Jorgensen n.d., Schmitt,Noe, Meritt and Fitzgerald, 1984, Tracy and Schuttenberg, 1989) or with otherissues such as effective criteria (Gips, 1985) and comparative cost effectiveness ofalternative approaches to selection (Hogan and Zenke, 1986). Finally, someauthors provided models of selection processes for school administrators basedon practical experience (Musella, 1983, Parkay, 1987). A particularly interestingapproach to selection was described in detail by Jantz, Hellervik and Gilmore(1986). These authors developed a technique called 'behavior-description inter-viewing'. Although designed for application in business and industry, this tech-nique has significant implications for selection of school-leaders. Jantz et al.provided a research-based argument for their set of procedures.


Based on available evidence, school-leader selection procedures are typicallymulti-step processes. A pool of aspirants become applicants and enter whatMusella (1983) called a 'decision-point funnel' with fewer and fewer individualsachieving candidate, and eventually 'selected' status. In the UK this process wasdescribed as long-listing, short-listing, preliminary interview, and final interview(Morgan et al., 1983).

There appeared to be considerable variation in selection processes acrossdistricts (Baltzell and Dentler, 1983, Schmitt and Schechtman, 1990). In Ontario,for example, the process ranged from a single interview with the chief educationofficer (a rather short funnel) to very elaborate, structured procedures that mightinclude mandatory participation in locally sponsored leadership courses or attend-ance at assessment centers (Musella and Lawton, 1986). Considerable variationacross districts also exists in terms of the decision mechanisms installed at eachcritical point in the funnel. Depending upon supply and demand, the majority ofapplicants were 'screened out' at very early stages of the process. Elkins (1987)reported that up to 80 percent of applicants were eliminated after the first screenin at least one LEA in the UK.

Musella and Lawton (1986) reported that written policies and procedureswere more prevalent in Ontario for principal and vice-principal selection than forany other role. Most such documents had been developed only recently, how-ever. In the UK, the written description of selection policies and procedures wasmuch less prevalent. Many of the processes observed were fairly unsystematicand non-explicit. There appeared to be a focus on individualism and 'felt need'supporting the selection of teachers for specific schools (Morgan et al., 1983).Elkins' (1987) data confirmed these findings. He argued that the typically consen-sual process of developing a profile of the ideal applicant was of questionablevalidity. Critcria for selection werc often obscure or only loosely defined incurrent practice. Frequently used critcria included biographical information,career track record, communication and interpersonal skills, knowledge of admin-istration, knowledge of the position, education/training, and appropriate phi-


2 ' (;

Selection Processes and Measurement Issues

losophy of education (Baltzell and Dent ler, 1983, Bryant et al., 1978, Elkins,1987, Morgan et al., 1983, Muse lla and Lawton, 1986, Schmitt and Schcchtman,1990). Many of these indicators have been criticized as being 'inevitably subjec-tive' (Morgan, et al., 1983, P. 58) and poor predictors of successful futureperformance (Jantz et al., 1986, Parkay, 1987). There is considerable support foruse of criteria that are more job relevant and directly linked to performance in therole. Indeed, the courts' insistence on such criteria have finessed professionaldeliberations on this matter. Several authors advocate the use of systematic andthorough position analysis as a way of identifying such criteria ( Jantz et al., 1986,Musala, 1983, Parkay, 1987). Analysis of this sort, it is claimed, would informthe interviewer about what to look for, ensure that all important dimensions ofperformance are covered, and reduce the likelihood that selection is based uponan unjustifiably narrow set of criteria (Parkay, 1987).

Assessment centers have become quite popular in the US, parts of Canada,and elsewhere. They .are used by many districts as part of the school-leaderselection process. One model of the assessment-center concept designed espe-cially for school-leaders has been developed by the National Association ofSecondary School Principals (NASSP). Versions of this model are now beingimplemented in over fifty locations in the US and Canada. Their features,advantages and disadvantages have been described in detail (e.g.. Allison, 1989,Gomez and Stephenson, 1987, Hersey, 1986, Milstein and Fiedler, 1989, Schmittet al., n.d., Schmitt et al., 1984). Participants' responses to a series of simulatedschool-administration tasks are evaluated by a panel of trained assessors. Princi-ples of performance-based evaluation and high objectivity lie at the heart of theassessment-center concept.

The tasks and assessment procedures associated with NASSP centers (seeFigure 14.1) were developed around a set of twelve criteria or dimensions ofperformance which emerged from an extensive job-analysis study conducted inthe United States. While job analysis is a systematic way of identifying criteria, ithas not gone without criticism. For example, the content validity of the NASSPassessment-center dimensions or criteria have been the subject of debate. Theresults of several studies appear to support claims that these criteria are validreflections of expert school leadership (e.g., Schmitt et al., 1984, Gomez andStephenson, 1987) but concerns persist, nevertheless. Allison (1989, p. 4), forexample, suggested that 'scholars inclined toward more global conceptions ofadministration ... might well take serious issue with the conceptual framework'.Bryant (1990, pp. 358-9) claimed that:

If the NASSP Assessment Center is to be trusted as an employmentscreen, it should be able to capture the accumulation of expertise that... is the logical by-product of the knowledge gained through variedand repetitive experience. The results of the investigator's analysis ...would suggest that the NASSP Assessment Center does not capture theassumed expertise (experience) possessed by participants on manydimensions.

Another concern regarding the NASSP dimensions of performance has to dowith the method used to derive them. The use of job-analysis technology entailsthe generation of a comprehensive list of relevant dimensions of performance.


Developing Expert Leadership for Future Schools

Figure 14.1: Assessment-center dimensions and descriptions

1 Problem analysis: Ability to seek out relevant data and analyze complex information todetermine the important elements of a problem situation; searching for information with apurpose.

2 Judgment: Skill in identifying educational needs and setting priorities; ability to reachlogical conclusions and make high-quality decisions based on available information; ability toevaluate critically written communications.

3. Organizational ability: Ability to plan, schedule, and control the work of others; skill inusing resources in an optimal fashion; ability to deal with a volume of paper work and heavydemands on one's time.

4. Decisiveness. Ability to recognize when a decision is required and to act quickly(without an assessment of the quality of the decision).

5 Leadership: Ability to recognize when a group requires direction, to get others involvedin solving problems, to interact effectively with a group, to guide them to theaccomplishment of a task

6 Sensitivity: Ability to perceive the needs, concerns, and personal problems of others,tact in dealing with persons from different backgrounds, skill in resolving conflicts, ability todeal effectively with people concerning emotional issues, knowing what information tocommunicate and to whom

7 Range of interests: Competence to discuss a variety of subjects (educational, political,economic, etc I, desire to actively participate in events.

8 Personal motivation Showing that work is important to personal satisfaction, a need toachieve in all activities attempted. ability to be self-policing

9 Educational values Possession of well-reasoned education, philosophy. receptivenessto change and new Ideas

10 Stress tolerance Ability to perform under pressure and opposition, ability to think onone's feet

11 Oral-communication skills: Ability to make a clear oral presentation of ideas and facts

12 Written-communication skills Ability to express ideas clearly in writing, to writeappropriately for different audiences such as students, teachers, parents, or otheradministrators

Source Schmitt et ai (1984) juumai ot Applied Psychology

which are then rated for importance by a representative sample of job incum-bents. A potential problem with this method lies in the sampling strategy. Failureto distinguish between expert and non-expert job incumbcnts means that ratingsof importance reflect the views of non-expert to expert school-leaders with theweight of opinion favoring the non-experts, since experts, as we have definedthem in this book, usually represent only a small proportion of the school-leaderpopulation. This was the case in the development of the NASSP framework. It isconsiderably more desirable to derive importance ratings from expert job incum-bents, especially when the purpose is the selection of future leaders. Results of thetype of research reported in earlier chapters of this book (especially Chapters 5



Selection Nocesses and Measurement Issues

and 6) provide, in our view, a more viable source of criteria for the selection offuture school-leaders. While the research directly supports the validity of some ofthe NASSP criteria appearing in Figure 14.1 (e.g., problem analysis) and rulesout very few, it also identifies criteria which do not appear in Figure 14.1 of arenot given adequate weight: for example, in the case of high-ground expertise,goals lor vision) and the extensive knowledge of curriculum and instructionsubsumed by factors.

Nonetheless, the NASSP criteria are the result of a systematic attempt tomove away from less well grounded criteria often used for selection purposes.Assessment centers are growing in popularity and recent evidence suggests thatthey are used for development purposes (e.g.. Milestein and Fiedler 1989, Tracyand Schuttenberg, 1989). Further, they are rarely the only source of evidenceused for the purpose of selection (Allison and Allison, 1990, Hersey, 1986).

Da(a Collet-non

Although a wide variety of data appear to be gathered for the purpose ofschool-leader selection, the interview is clearly the most popular source of in-tbrmation for making decisions about candidates (Baltzell and Dentler. 1983,Bryant et al., 1978, Morgan et al., 1983, Musella and Lawton, 1986, Schmitt andSchechtman, 1990). Interviews were found to be used quite extensively at bothshort-listing and final-decision points. While many districts undoubtedly utilizestructured interview questions that are applied to all candidates, open-ended orunstructured interviews are still fairly common in practice. Morgan et al. (1983)observed a considerable amount of unstructured interviewing in their study in theUK. Baltzell and Dentler (1983) also reported substantial variation in interviewstructure over school districts in the US.

The dangers in using unstructured interviews have been well documented( jantz et al., 1986. Morgan et al., 1983). A tendency to rely on subjective criteriaand to collect data that are not comparable across candidates are the two chiefconcerns. The use of structured interviews, as a minimum, helps to ensure thatcomparable data ;.re collected.

Next to personal interviews, resumes and references appeared to be thesources of information with the highest perceived value to selectors (Baltzell andDentler, 1983, Musella and Lawton, 1986, Schmitt and Schechtman, 1990).Other sources included application forms (usually requiring education, experi-ence, and certification information), written statements by candidates about theirpersonal philosophy of education, and schedules and summary sheets. Bryant etal. (1978) repoved that superintendents ascribed importance to neat, grammatic-al, business-like letters of application.

No research evidence was available about current practice to suggest thatstandardized instruments were being used to measure potential expertise in theschool-leader role. However, several such instruments have been developed inrecent years (e.g.. Cousins and Leithwood, 1987a, Hallinger and Murphy, 1985,Ellett, 1978, Vandenberghe, 1988). While many of these (often behavior ratingscales) instruments have been designed specifically for appraisal (and research)purposes, they offer promise as screening mechanisms in school-leader selection(more on this in the next section of this chapter).


Developing Expert Leadership for Future Schools

In NASSP assessment centers, assessors observe a small number of candi-dates performing various tasks under very intense circumstances, usually forabout a two-day period. Then, individually and collectively, assessors reviewdata for each candidate and develop a consensus about his or her performance.Candidates are assigned scores on each of the twelve performance dimensions anda summary score is calculated. Although not as high as for assessment centers inbusiness and industry, 'inter-rater' reliability for performance judgments wasreported to be quite high (Schmitt et al., n.d.).

Who is involved in the selection process? There is considerable variationwithin and across districts regarding those responsible for selection, variationwhich depends upon the poine in the decision-making funnel. Musella andLawton (1986) observed that prior to having one's application for a formalschool-leadership position taken seriously, a favorable recommendation from theapplicant's current principal must have been submitted. They also noted that insonic cases 'pre-consideration interviews' might be held. From that point, selec-tion committees consisting of district administrators, current school-leaders, trus-tees and even CEOs assumed responsibility for coordinating the selectionprocess. Final dccisions about candidates were most often made by electedofficials or by the CEO with trustee approval.

Evidence from US research suggested that the final decision also rested withthe CEO, with rather unstructured screening processes being conducted bypersonnel directors and senior assistants to the CEO (13altzell and Dentler, 1983,Bryant, et al., 1978, Schmitt and Schechtman, 1990). Similarly, in the UK,elected LEA members, governors, and advisory officers were typically involvedin the selection process (Morgan et al., 1983). The extent to which advisoryofficers had selection as part of their portfolio was found to vary across systems.Different selector groups were found to be active at different 'stages of elimina-tion'. Sometimes this resulted in a completely new group of selectors beingpresent for the final phases of the selection process (Elkins, 1987, Morgan et al.,1983). Candidates might have been required to face a panel of up to thirtyselectors. Generally, CEOs and governors assumed responsibility for final deci-sions about head-teachers' appointments.

in the case of assessment centers, assessors were typically district administra-tors from the school systems that were active members of the centre. Indeed,such contribution of district-administrator time was usually a requirement ofmembership. This requirement helped to ensure that assessors were never re-sponsible for assessing someone from their own district, thus reducing varioustypes of potential bias in the process. The extent to which these data were usedfor selection was quite variable, as well as who at the system level was responsi-ble for evaluating the information (Allison and Allison. 1990). As noted earlier,when assessment-center data were used for selection purposes. such data wereamong a larger set collected from other sources too.

Reporung and tollou,-up

We have noted that selection decisions were usually made by trustees, CEOs andothcr district administrators. Procedures for reporting such decisions were


2 5 0

Selection Processes and Measurement Issues

varied, depending upon the needs of the district and the decision process in place.Personnel files, with accumulated bits of information about candidates and some-times summary sheets, were used to support the final decision-making process.The extent to which these documents convey the most useful information aboutcandidates can be limited. This was especially the case where those making thefinal selection decisions did not participate in earlier stages of data collection anddecision-making (Morgan et al., 1983).

Jantz et al. (1986) called for the systematization of note-taking during theselection process. They recommend that notes be reviewed and clarified aftereach interview and that some measure of standardization over candidates ispracticed. With the exception of assessment centers, we were unable to findevidence of such practice in the currcnt literature on school-leader selection.Following the assessment process, assessors typically reached agreement on acandidate's performance and a standardized summary report was prepared. Thereport was shared with the candidate and with the sponsoring school-systemofficials. Muse Ila (1983) advocated the use of training in the selection process atthree different levels. First, training should be provided for those aspiring to therole. Second, training or debriefing should be provided for those who have beenscreened out of the process so that they can work toward developing the requiredqualities for future consideration. Finally, training should be provided on anongoing basis for candidates once they are selected. Parkay and Currie (1989)provided evidence revealing that principals perceived a gap in the availability ofsuch training opportunities. The extent to which such training occurred is un-clear. Muse lla and Lawton (1986), on the other hand, found that several Ontariodistricts have their own training programs designed to develop leadership amongtheir own staffs. The quality of these programs, however, remains unclear. Somedistricts sent candidates to assessment centers for the purpose of professionaldevelopment, as reported earlier. Also, as described in Chapter 11, some regionsin Canada requii ed certification for the principalship prior to securing such aposition. On the other hand, Elkins (1987) found that very few candidates forhead-teacher ever began their teaching career with such a goal in mind and thatteachers usually had very little knowledge about what heads did, let alone how toprepare for the role. Training opportunities appeared to be very limited for thoseaspiring to thc role of head-teacher in the UK. Current research revealed thatvery little follow-up of this nature was occurring in practice.

Evolution of Policy

There was little evidence about how selection policies were developed, whoparticipated in their development and whether they were subject to review andrevision. As mentioned earlier, Muse lla and Lawton (1986) found that docu-mentation for school-leader selection was more prevalent than for any other rolein Ontario. With respcct to the UK, we were unable to find evidence of changesince the rather dismal picture provided by Morgan et a/. (1983). There appearedto be a strong need to integrate, into existing policy and practice, a reviewmechanism consciously to examine the utility and effectiveness of school-leaderselection.



Developing Expert Leadership for Future Schools


Two questions are central to considerations about the impact of selection proce-dures: 'To what extent do data derived from the selection process support thedecisions made?' and 'To what extent does the information gathered and used fordecision-making predict success in the role at a later point in time?'

On the first question, there was substantial variation depending upon thesophistication of the process used. Where criteria were not well specified and datawere not systematically processed, the influences of local political processes,patronage, favoritism, and individual advocacy remained unchecked (Baltzell andDent ler, 1983, Elkins, 1987, Morgan et al., 1983). Where use was made of moreexplicit processes with multiple decision points, checks, balances, and provisionsfor data integration, it was likely that more objective, fair, and valid decisionswere going to result (Muse lla, 1983, Parkay, 1987). Inasmuch as non-job-relatedcriteria are in frequent use, we can assume that the quality of data for selectionpurposes often will predict future performance on the job very poorly indeed.Elkins (1987) reported this to be the case espec-ially for the early screening stageof the selection process. Jantz et al. (1986) described the inadequacy of suchcriteria as biographical facts, technical knowledge, experience and activity de-scriptions, and self-evaluative statements. They claimed that grounding criteria inthe performance base of candidates will improve substantially the predictivevalidity of selection data.

As we indicated above, the predictive validity of NASSP assessment-centerdata has been the object of considerable debate. While some concluded that thepredictive validity of the centers has been demonstrated to be satisfactory (Gomezand Stephenson, 1987, Hersey, 1986, Schmitt, et al., n.d., Schmitt et al., 1984),others remained skeptical (Allison, 1989, Bryant, 1990) However, an additionalchief concern corresponding to the utility of assessment centers for selectionpurposes has to do with costs (Milestein and Fiedler, 1989). Costs to individualsystems were substantial, given fees for candidates and the time requirements ofproviding assessors. Hogan and Zenke (1986) conducted a cost-benefit analysisthat favored the use of assessment centers and performance-assessment tasks forselection purposes as compared to traditional interviews and paper and pencilinventories. The estimate of predictive validity they used in their formula seemssubstantially higher than estimates reported elsewhere, however. Allison andAllison (1990) reported that school systems were not often in a fiscal position tosend all potential candidates for selection to assessment centers and, as a conse-quence, the use of such data for selection purposes became highly limited. Thedevelopmental benefits of assessment centers may outweigh their use as aids tothe selection process. Clearly, reports from candidates (Allison and Allison, 1989,Nagy and Allison, 1988, Tracy and Schuttenberg, 1989) and personnel of schoolsystems sponsoring candidates (Allison and Allison, 199(1) tended to support thisclaim. As vet. there was limited systematic evidence to support this claim (e.g.,Tracy and Schuttenberg, 1989).


The review of evidence concerning current school-leader selection processespaints a bleak picture of such processes. Under typical circumstances, for exam-



Selection Processes and Measurement Issues

pie, selection processes may indade the use of mappropriate, ambiguous and/orinvalid criteria, and the collection of data which arc both an unreliable and invalidreflection of such criteria. Assessment-center processes, especially those devel-oped by NASSP, offer promising responses to the limitations of typicalselection processes. In practice, however, assessment centers are subject to critic-isms, especially regarding their choice of criteria and the predictive validity oftheir results. Perhaps of even more consequence is their cost, which is clearlyprohibitive to many districts.

While the review of current selection practices in this section of the chaptercontains a number of implications for improvement, much work needs to bedone before se!ection processes can be relied on to contribute to the developmentof leadership expertise in future schools. At the heart of much of this work are aseries of issues regarding the measurement of school-leader expertise: theseissues, of course, are also important to performance-appraisal processes.

Measuring School-Leader Expertise

In this section, current approaches to measuring school-leader expertise are de-scribed. This is followed by an analysis of several especially critical measurementissues. We then examine inn more detail the measurement of expertise on the highground. Research to validate one instrument, the Principal Profile-Based Instru-ment (PPBI), is summarized briefly and the advantages of the instrument arenoted. Finally, an agenda for further development work in this area is outlined.Special attention is given to the challenges of assessing expertise in the swamp.Those readers less interested in the inherently technical issucs explored here maywish to skip directly to 'Implications for Improvement'.

Instruments for Measuring School-Leader Expertise

Ginsberg and Berry (I 990) identified instruments and approaches that have beenused in the selection and appraisal of school-leaders. Undoubtedly their samplewas incomplete because thousands of school districts have developed their owncheck-lists and inventories and only a fraction of these are reported in theresearch or professional literature.

Our search for existing instruments was guided by two criteria. The reporthad to:

show that the instrument was either designed for, or adaptable to theproblem of school-leader assessmentdescribe the instrument in terms of us

formatorigins and/or development, andpsychometric properties (however meager the description).

Using these criteria, fourteen instruments, aspects of which arc described inTable 14.1, were identified primarily through a computerized (ERIC) search andbibliographic follow-up. While this set of fourteen is clearly not exhaustive, it is


2 5 3



Table 14 1 instrument for measuring school-leader expertise

Instrument Description Validity Reliability Comments

Andrews. Soderand Jacoby 119861

(Tous:ns andLeithwood (1987a1Cousins (1989)

1. Staff-AssessmentQuestionnaire(SAO)

19 items from 167-item questionnaire5 point scale

2 Pfiricipai 68 Items BARSBased 4 subscalesInstrument 6 point scale(PPRI)

Prneier and Wdson 3 School-Principal(19891 Effectiveness-


1 W.', i1918; 4 PrincipaiPerformanceDes ^I pt,ry,


i4acter and MJrol-,v 5 Principali1985 19871 in,,tructional


Rating Scaif,IPIMPIS


5 questionnaires(student, staff.parent, principal,supervisor)213 items

- S and 6 point scales

4 questionnairesiprincipal, teachereoernal. ohserypr,supervisor)100 items

71 items BARS- 11 subscales

b pulid

Based on review of Alpha = 0 93instructional (High)leadershipliterature(Moderate)

Based on PrincipalProfileEffectivenessresearch,literature review,empirical validation(Moderate High)

Based onconceptualframework andsubjectivedetermination ofitems(Moderate Low)

Based on extensivecontent-validationprocedure lobanalysis(Moderate)

Based on review ofliterature oninstiuctiundileadershipempiricalvalidation(High)

Alpha = 0 98Subscales0 91 to 0 96(High)

Protected tobe 0 80+(Inconclusive)

Not reported

Alpha = 0 75+(Moderate High)


Requires furtherempiricalvalidation

development stage c;(5

Still in early

2 5 5

on() ars, Pc-4o,mance119h5) Review Analysis

andImprovementSvstem forEducationPRAISE I

K," ' t',3



3 0.-,3-.333





9 P. a Pro`,e

'J A

81 .teins9 subscales5 point scale

34 .tems- 5 builit sca,e

34 items3 5.,:bscaes

- 5 point scale

Multidimensional,multi-stagedriescriptio-, ofgrowth in oractce,nterviewassessmenttechrigue

Pe'formance-do:Jo.,montmstrument 'Orrat.ng resDOnCe rOProblem 3oktnqtasks4 point scaiehohstic ratogsby 2 erf, 1.4

Based on review ofeffectivenessliterature andextensive content-validation procedure(High)

Based on Powertheoreticalframework of Spadyand Mitchell 0977)(Inconclusive)

Based on review of!:terature andempirical validation(Moderate High)

Based on review ofliterature andempirical .,alidation(Moderate High)

Showed past.ntet4ent.on gar. .nProblem-solvingsk.II(InconclusiveI

Alpha = 0 88 to 0 98Test-retest0 59 to 0 80(Moderate High)

Test-retest andexpert-novice datareported in non-traditional manner(!nconc)usivel

Alpha = 0 7 to 0 9(Moderate High)

Inter-rater agreement0 50 to 1 00(Moderate High)

for use in analyzingresponses toperformance-assessment tasks

tor use primarily asa diagnostic tool

Inter-rater agreement for use in analyzingto 95% responses to

(Moderate High) Performance-assessment tasks


Table 14 1 (cont

Sourcels) !nstiument Description Validity Reliability Comments

P.tner and Hocevar(1987)

Tucker and Bray(1986)

V&entine aridHow--1,3r

Vandenbe.ghe (19'38)

11 Yuki's 115 itemsManagement- 23 subscalesBehavior Survey 5 point scale

12 Profile for theAssessment otLeaders(PAi

13 Audit of PrincloalE tfectiveness

14 2.ange-f- ar.,litator StyleOuestionnaveiCr SO)

88 Items8 sirbscales

80 items3 domains9 subscales

-71 items3 domains6 subscalc s6 r,o.nt scale

9 subscaleseliminated due tonon-responseconfirmatory factorvalidity established(Moderate)

Items drawn frominitial list of 10.000+empirical validation;teachersprincipals)(Moderate)

Review of literature(not cited,factor validity(Inconclusive)

Alpha = 0 78 to 0 93(Moderate High)

Not reported

Alpha = 0 79 to 0 93(Moderate High)

Based on policy Alpha = 0 64 to 0 95implementation (Moderate)literatureempirical vahdationfactor validity(Moderate High)


Insti ument adapted ument adaptedfrom business and business and,ndustry stry

Items initially drawnfrom behaviorsidentified by ProtectROME(Ellett, 1978)

Is initially drawnbehaviors

titled by ProtectAE

tt, 1978)

Selection Processes and Measurement Issues

probably sufficiently represemative of the field to sustain our subsequentobservations.

All fourteen of the instruments examined have been developed in recentyears. The Principal Performance Description Survey (PPDS) reported upon byEllett (1978) was the earliest to be developed of the instruments (4 in Table 14.1)that came to our attention. The majority of the instruments were rating scalesand most had versions designed for self-ratings and for ratings of school-leadersby others such as teachers, superordinates, parents and students. The ratings aregenerally `Lickere type scales ranging from 4 to 6 poinis. Some instruments takethe form of behavior-observation scales (BOS) (e.g., Cousins and Leithwood,1987a, Hollinger and Murphy, 1985; 2 and 5 in Table 14.1) while others includedratings of attitudes, competencies or knowledge. Two instruments (Krysinski,Reed, Gougeon and Armstrong, 1987, Leithwood and Steinbach, 1990; 7 and 10)were designed specifically for the purpose of helping to assess the performance ofschool-leaders on simulated tasks.

There was considerable variation in the length and comprehensiveness ofthe instruments. Some were quite long (e.g., 115 items, Ellett, 1978, 100 items,Pitner and Hocevar, 1987; 4 and 11) but the average length was between sixty toninety items. Others consisted of a battery of questionnaires designed for com-pletion by different groups of raters (e.g.. Ebmeier and Wilson, 1989, Ellett,1978; 3 and 4) Most also had several subscales corresponding to dimensions orcomponents of school-leader expertise.

Validity data were highly variable. At least two different approaches to thequestion of content validity were employed. Some instrument developers focusedon effective school-leader practices identified through reviews of effectivenessliterature (e.g., Andrews, Soder and Jacoby, 1986; I) or empirical validationemploying panels of 'experts' as judges of content (e.g., Cousins and Lcithwood,1987a, Larsen, 1987; 2 and 8). Other developers derived sets of items fromreviews of research and obtained ratings of importance from representative sam-ples of practitioners including supervisors, principals, and teachers (e.g., Ellett,1978, Hollinger and Murphy, 1985, Knoop and Common, 1985; 4, 5 and 6). Yetothers relied on statistical validation usually employing exploratory factor analy-sis (e.g., Knoop and Common, 1985 Valentine and Bowman, 1986, Vanden-berghe, 1988; 6, 13 and 14) to make decisions about how variables clusteredtogether. Pitner and Hoccvar (1987; 11) used confirmatory factor analysis, aprocedure that requires a priori decisions about factor structure and variableloadings. Some developers weie less rigorous in their pursuit of validity evidenceand were content to generate and informally field-test items based upon aconceptual framework (e.g., Ebmeier and Wilson, 1989; 3). Several studiesemployed tests of different types of validity including content, construct anddiscriminant validities (Cousins, 1989, Hollinger and Murphy, 1985; 2 and 5).

The reliability of these instruments was generally found to be quite good,but often this could be attributed to the large number of items and raters in thesample. Internal consistency using Cronbach's Alpha was the most popular testof reliability. Coefficients ranged from about the 0.70 to 0.95+ range with somevariation over subscales. In some cases, reliability data were not reported (e.gEllett, 1978, Tucker and Bray, 1986; 4 and 12) or were reported in a casual

manner (Krysinski et al., 1987; 7). In other cases, test-retest reliability wasreported (e.g.. Knoop and Common, 1985; 6). None of tho reports cited inter-rater reliability analyses.


Developing Expert Leadership for Future Schools

Most developers claimea multiple purposes for their instruments such asresearch, appraisal, and selection. Some instruments were designed specificallyfor the purpose of providing diagnostic information for growth (Leithwood andMontgomery, 1986, Leithwood, 1987; 9). Other instruments were designed asaids for the analysis of performance data (e.g.. Krysmski et al., 1987. Leithwoodand Steinbach, 1990; 7 and 10).

Key Issues in the Measurement of School-Leader Expertise

While there appear to be a number of promising instruments to help in theselection and appraisal of school-leaders, careful attention needs to be devoted tosome of the limitations and potential dangers in such use.

The Joint Committee on Standards for Educational Evaluation has recentlypublished a set of standards fOr personnel evaluation practice ( Joint Committee,1988). Prior to this publication, we participated in the public hearings' processcoordinated by the joint Committee (Cousins and Leithwood, 19876). Thisparticipation gave us the opportunity to compare a pre-publication version of thestandards against, among other things, the knowledge-utilization frameworkdescribed in Chapter 13 (see Figure 13.1). Figure 14.2 summarizes the results ofour comparison of the standards against our knowledge-utilization frameworkand demonstrates a high level of consistency between the two.'

Most of the standards apply to and can be categorized according to theappraisal-implementation dimension of the knowledge-utilization framework.Also, most are associated with factors that have a direct connection to measure-ment issues, specifically accuracy standards such as valid measurement, reliablemeasurement, systematic data control, sufficient sample of performance (thisstandard has since been deleted by the Joint Committee), biased control, anddefined role. Several of the utility standards, further, are associated with factorsfrom the implementation dimension of the knowledge-utilization frameworkincluding defined uses, valuational interpretation, and evaluator credibility.

However terse, this analysis supports the usc of thc standards to assist inimproving the measurement of school-leader expertise, improving not only re-liability and validity but also impact (use). But the task of measuring school-leader expertise is complex. and 'measuring up' to the standards set by the JointCommittee will not be easy. We turn now to a more dircct examination of somekey issues involved in achieving such a goal.

D. and specification of what is to be measuredMi...,..iy (1988) provided a comprehensive analysis of problems involved inmeasuring instructional leadership. Many of the instruments that were describedearlier, as well as the criteria employed by the NASSP assessment centers werebased on job-analysis techniques, an approach that Murphy considered to beflawed for two rcasons. First, the results of job analysis often do not provide cleardefinition of the qualities required of the job. Second. instruments based on jobanalysis are unrelated to any coherent theory of expert practice. This secondproblem is likely to he the more serious. Unclear definition of job qualitiesinfluences instrument reliability but the reliability of at least the instrumentswhich were reviewed earlier in this chapter was quite high. Lack of a coherent

2 61

Selection Processes and Measurement Issues

Figure 14 2. Personnel evaluation standards by knowledge-utilization factors

Evaluation-implementation Factors11 Evaluation quality

21 Cred.b.'

3) Relevance

4) Communication quality5) Timeliness

61 Nature of results (Findingsi

Decision or Policy-setting factors1) information needs2) Significance of decision

(Decision characteristics)3) Political climate

4) Competing information5) Personal characteristics61 User commitment

Appraiser/evaluation-subject dynamics

F-1 Practical procedures4-4 Valid measurementA-5 Reliable measurement4-6 Systematic data controlU-3 Evaluator credibility

Valuational interpretation* Sufficient sample of performanceA-7 Biased controlU-2 Defined uses

Valuational interpretationA-1 Defined roleU-4 Functional reportingU-4 Functional reportingU-5 Follow-up and impactU-1 Constructive orientationU-5 Follow-up and impact

U-2 Defined usesNone

F-2 Political viabilityP-3 Conflict of interest4-2 Work environmentNoneNone

F-2 Political viabilityP-5 Interaction with subjects of evaluation

Note From previous draft of The StandardsSource. Adapted from Cousins and Leithwood (1987b) JCSEV, Washington, UL,

theory, however, calls into question the construct validity of instruments.According to Murphy, the job-analysis approach yields lists of variables that aredescriptive of 'typical' practice. If there is a need to measure expert practice, aconception of such expertise is needed to generate items for the instrument.Hence, instrument development will need to include, as part of the process ofreviews, research on expertise and/or empirical validation using panels of ex-perts. A coherent conception of expert school leadership such as that provided inChapters 5, 6 and 7 should form the basis of any instrument to be used for thepurpose of measuring school-leadership expertise.

AssessmentMurphy (1988) also expressed concerns about the use of behavior-based instru-ments in isolation. Results of such instruments alone ignore thc contextualrealities of schools. Organizational conditions, symbolic and cultural activities,and established organintional routines are important parameters of the school-leader's role which should be considered in concert with instrument data. Failureto do so may give a distorted impression of a school-leader's true expertise.

Two other issues concerning assessment arc also worth noting. First, as Hazi


Developing Expert Leadership 1-or Future Schools

(1989) reminded us, there is a tendency to trivialize our conception of perform-ance in complex roles and to fall prey to the illusion that instruments are errorfree. Instruments do involve human judgment and subjectivity does enter intothe processes of administering an instrument. This underscores the need to collectmultiple types of data from multiple sources. No single measure of a school-leader's expertise is likely to be a truly valid picture of his or her capabilities.Second, and more fundamentally, it is necessary to consider those aspects ofschool leadership that do and do not lend themselves to measurement. Consider,for example, aspects of expertise on the high ground. Strategies used by school-leaders and elements of their decision-making are more directly suited tomeasurement with the kinds of instruments that are currently available. Butwhen the focus shifts to other aspects of expertise the nature and role of values,for example what are the implications for measurement? A call for a funda-mental shift in our approach to the measurement problem, one that adheres toperformance-assessment technologies is supported.

Potential fir misuseCousins. Begley and Leithwood (1988) have explored the question of potentialmisuse of instruments for the purposes of school-leader selection and appraisal.Given the information needs of decision makers and the need of school-leadersfor profrssional development, what are the possible or probable misuses ofavailable instruments? A useful framework for responding to this question wasprovided by Alkin and Coyle (1988). Patton (1988), (cited by Alkm and Coyle)identified two dimensions relevant to our analysis of this question: utilization andmisutilization. Utilization of performance data varies from use for decision-making or conceptual development (depending on purpose) at one extreme andnon-utilization (e.g., failure for users to process data) at the other. Similarly,variation in misuse may be plotted along a continuum ranging from misutiliza-tion to justifiable action.' As a 1 aristic for differentiating types of misuse, itmight be helpful to consider these dimensions to be completely independent ofone another as shown in Figure 14.3.

Alkin and Coyie (1988) described several distinct variations of misutilization.Data derived from an instrument may be expected to vary in technical quality. Itmay be more or less free of measurement error (reliable), and derived frommultiple raters such as school-leaders, district administrators, and teachers(valid). If information about an individual is known to be of superior technicalquality but is suppressed by a potential user (e.g., selection-committee member)for whatever reason, we have an instance of what Alkin and Coyle refer to asabuse: a clear case of non-utilization of the data that can readily be described asmisutilization.

But some instances of non-utilization may be legitimate. For example, if aselector was aware that the results of an assessment were technically flawed orerroneous, he or she would be justified in not incorporating the information intothe selection process. This would constitute an example of justified non-use, anappropriate and responsible action. If, on the other hand, the data are of sufficienttechnical quality but potential users are unaware of their existence, or inadver-tently fail to process the information, this is a case of unintentional non-use. Whilethis is not misutilization, it is certainly not a desirable outcome.

The best possible outcome is one in which goo-I-quality instrument data arc



Selection Processes and Measurement Issues

Figure 14 3 Conception of misutilization




MsevaluationIdeal Use

Justifiable Action



Umntent,onal Non-use

Justified Non-use l!ntentionail

Non-u tiliza (ion

Source Adapted from Akin and Coyle (1988) Studies in Educational Evaluabon

thought about by users and subsequently lead to further conceptual developmentof the users and/or assist them in making specific decisions. This would be use inthe ideal. A less satisfactory, but nonetheless legitimate form of use, would be forpotential users to actively think about the meaning of data provided through aninstrument but, subsequently, not to learn from them nor base decisions uponthem. Such use can be quite legitimate. For example, a school-leader might fullyunderstand the information provided by an instrument used in his or her apprais-al but consider it 'old news'. There is nothing new to be learned thai would behelpful in fostering further self development.

Finally, Alkin and Coyle differentiated between two types of misutilizationwhcn data are processed. The first they described as misuse, a term that corres-ponds to the deliberate manipulation of, say, instrument scores to serve someparticular purpose (e.g., support or non-support for a school-leader candidate).Clearly, this situation is an example of intentional misutilization of the data, sincethe data are used in an inappropriate fashion. The second type of misutilization ismisevaluation. One type of misevaluation occurs when test developers do not takethe necessary steps to prevent misutilization. Incomplete scoring information,absence of normative data, poor administrative instructions, and the like, arepossible sources of error that could ultimately lead to misevaluation. Of course,the responsibility for misevaluation need not necessarily reside with the instru-ment developer. Misevaluation could be the result of careless instrument admin-istration, scoring, and the like. Cousins ei al. (1988) identified a number of

2 6 12.37

Developing Expert Leadership for Future Schools

additional, potential misuses of data about school-leader expertise. The examplesalready provided, however, are sufficient to alert the reader to the possibilities ofmisuse. The framework captured in Figure 14.3 provides a means for becomingmore sensitive to misuses that might be encountered in one's own context.

The final section of this chapter touches on ways of avoiding potentialmisuses of data about school-leader expertise. Having examined issues whichappear especially noteworthy concerning the measurement of school-leader ex-pertise, we now turn to the development of an instrument designed specificallyto measure school-leader expertise on the high ground. In doing so we willdemonstrate how at least some of the issues discussed above have been addressed.

Measuring Expertise on the High Ground

The Principal Profile-Based Instrument (PPBI) was initially developed andrefined by Cousins and Lcithwood (1987a). Items for the PPBI were deriveddirectly from the conception of high-ground expertise described in Chapter 5. Aneighty-item version of the instrument was subjected to a content-validation studywith a panel of 'expert' district administrators and principals serving as judges.Thc refined version included sixty-eight items; self-report and colleague forms ofthe instrument were produced for both incumbent principals and those aspiringto the role, respectively. The self-report form of the PPBI for incumbentsappears in Appendix A. We consider this version of the instrument to be 'contentvalid': that is, the items were judged to be representative of high-ground school-leader expertise (Cousins and Leithwood, 1987a). Additional evidence supportingthe reliability and validity of the instrument was collected in a study by Cousins(1989). Specifically, this study aimed at following objectives concerning thePPBI: (1) to establish a central data base for the PPBI; (2) to examine normativeresponse distributions of the instrument for both principals and colleagues; and(3) to provide further evidence about the validity and reliability of the instru-ment. Results of this study are reasonably encouraging from a psychometricpoint of view. The reliability of the instrument was found to be very high andthere was some (less convincing) support for its construct validity. These are notconclusive analyses, however, and several limitations should be kept in mind.

First, the sample upon which these analyses were based was selected out ofconvenience. As such, the external validity of thc results may be questioned.Clearly, further work is required to collect information from a sample of school-leaders inure demonstrably representative of their colleagues. This step precedesany confidence in 'norming' the PPI31. Second, other aspects concerning validityshould be examined. Specifically, more rigorous and sophisticated tests of predic-tive validity and concurrent validity should be conducted. Also, a more directtest of the extent to which subscale scores do measure different constructs (factorvalidity) would provide valuable information concerning the PPI31.

The PPBI is potentially useful as part of a set of school-leader selection orappraisal procedures (or as a tool for school-lcadcr research). Our studies to dateprovide encouraging results concerning the psychometric properties of the instru-inent. Further work seems likely to yield useful results for measuring school-leader expertise on the high ground.


Selection Processes and Measurement Issues

Measuring Expertise in thf Swamp: Challenges Ahead

Accurately measuring expertise on the high ground is a task which has beenseriously underestimated in school districts. Nonetheless, as we have demons-trated, research about expertise on the high ground is sufficiently well developedto be quite informative to the measurement process. Measuring expertise in theswamp has received much less attention to date and, for this reason, we arefarther from being able to provide dependable, well-tried advice. Nevertheless,Murphy (1988, p. 122) raised some concerns that apply to the measurement ofexpertise in the swamp such as for example, his criticisms of job analysis de-scribed above, and a focus on:

behaviors that are directly observable, technical in nature, directly linkedto curriculum and instruction, visible in the short term and close inproximity to the consequent action or effect.

While many of the qualities associated with high-ground expertise are overlookedby such approaches, this is even more the case in relation to expertise in theswamp. A clear advantage to the use of behavior-rating scales, however, is 'theconceptual clarity with which the performance domain is defined' (Borman.P)86. p. 113). But, as we have argued, such conceptual clarity may not pertain toall aspects of the role of future school leaders.

How then can one go about the business of assessing expertise in theswamp? Our response to this question is limited to the answers which have beendeveloped in the course of our research program. In that context, one instrumenthas been developed which offers the beginnings of a solution to the measurementproblem. This instrument (broadly defined) is based on observing school-leaders'responses to simulated job tasks. Such simulations are, in fact, the cornerstone ofthe NASSP assessment center described earlier in this chapter. Also outlined werethe elaborate procedures employed by the centers and how relatively objectivemeasures of performance are obtained. The basic technology for measurement ofthis nature is described by Stiggins (1987) in his instructional module titled'Design and development of performance assessment'. It is this approach that werecommend as the beginnings of an approach to measurement in the swamp.

The fundamental principle of performance assessment is that measurement isbased on professional observation and judgment. It calls for the use of trainedraters and when used carefully can produce reliable and valid results. As Stiggins(1987, p. 33) suggested, `the keys to success ... are (a) to make the judgment-based evaluation as systcmatic and objective as it can be while (b) focusing on thernost important attributes of performance'.

Performance-assessment techniques are useful where the assessor is in-terested in measuring the translation of knowledge and understanding intoaction. They typically involve a written response to a natural event (problem)framing the kind of performance required. Scoring is usually done by more thanone rater and then ratings are compared for agreement. In the NASSP model, forexample, several raters reach a consensus agreement about scores for a particularassessment-candidate. Data provided by Schmitt et al. (n.d.) indicated that inter-rater reliability is usually quite high for these ratings. Perhaps the main reason for

266 239

Developing Expert Leadership for Future Schools

this outcome is the extensi e training that assessors undergo. This is an essentialelement of performance-assessment technology.

As Stiggins (1987) indicated, a major advantage of performance assessment isthat it provides rich evidence of the use of skills and knowledge one is interestedin assessing. The technology is performance-based and as such transcendsthe simple regurgitation of knowledge about how one might perform. This, thereader will recall, is a key feature of behavior-description interviewing describedpreviously (jantz et al., 1986).

An additional requirement of performance assessment is that criteria arespecified precisely and that they can be communicated to raters. This calls for theuse of a valid framework to guide the scoring. In the NASSP model and the'behavior-description interviewing' approach, the frameworks are derived fromjob analysis but criteria need not be restricted to this source. Our own scoringprocedures were derived from our research-based framework for school-leaderexpertise in the swamp described in Chapter 6. A bit more detail about ourinstrument will help to demonstrate the application of performance-assessmenttechnology to th,- problem of measuring school-leader expertise in the swamp.

In the study alluded to in Chapter 12, we had experimental and controlschool-leaders provide written solutions (called 'protocols') to four hypotheticalcase problems. Respondents received different problems for which they wrotesolutions before and after the program was conducted. These written protocolswere rated holistically for the quality of both the problem-solving process usedand the product or solution generated. Protocols were also analyzed for evidenceof the strategies explicitly taught in the experimental program. Two expert raterswere trained to rate (on a 4-point scale) the protocols for the quality of bothprocess and product. Raters did not know which of the two groups the protocolswere from or whether a protocol was generated before or after the experimentalprogram was conducted. The scoring yielded overall results as well as ratings ofindividual components of the problem-solving framework (e.g., interpretation,goals, values, etc.). This illustrates the application of performance-assessmenttechnology to the measurement of expertise in the swamp. As noted by Stiggins(1987), the keys to success are: carefully prepared performance exercises; clearperformance expectations; and careful, thoughtful rating (something that requiresa fair bit of time).

Our own work in this area is in the early stages but we are encouraged bythe results, so far. Of course, the joint Committee's Standards apply to measure-ment of expertise in the swamp as well as to measurement of expertise on thehigh ground. Depending on training, reliability can reach sufficient levels; valid-ity is heavily dependent upon the underlying theory or framework used ingenerating performance criteria. 'Feasibility' is another standard that must beconsidered, especially in the context of selection and appraisal. Certainly, a

significant investment must be made in the training of raters and the time theyrequire for their tasks. Given the potential contribution of appraisal and selectionto developing expertise in the swamp, these arc resources well spent.

Summary and Implications for Improvement

Both appraisal and selection processes result in judgments: judgments aboutneeded growth, promotion. dismissal, transfer, and the like. Even in the case of



Selection Processes and Measurement Issues

appraisal for growth, school-leaders need to reflect upon, and make judgmcntsabout, their own performance. The adequacy of these judgments depends, criti-cally, on the quality of the information available to those making judgments andthe use of that information. This chapter and Chapter 13 have been preoccupiedwith both of these matters. But the most unique aspect of these chapters is theiremphasis on the use of information. This emphasis has emerged from our owngrowing awareness that a great deal of personnel-evaluation information appears

not to be used. Lack of use of such information means not only that the

significant resources devoted to its collection are wasted but also that there is nopossibility for such information to contribute to school-leader development our

main concern in this book.The guidelines outlined in this section for improving appraisal and selection

processes reflect our preoccupation with increasing the use of information col-

lected as part of these processes. These guidelioes are organized around the sameframework used to dcscribe current practices throughout these two chapters.Appraisal and selection processes are treated as one kind unless otherwiseindicated.


Specify policies and procedures:Selection and appraisal policies and procedures should be well specified,

documented, and available for consumption by all interested parties. Specificationshould include criteria and expectations, process, timeliness, appeal procedures,

and so forth. Selection policies should be multi-step with screening mechanismsand/or a decision-point funnel should be installed. Responsibilities ;;,scribed at

each decision point should be clearly outlined.Differentiate purposes and emphasize growth:

Purposes for selection and appraisal should be clearly understood by both eva-luators and those being assessed. While it is possible conceptually to separate the

two fundamental purposes (to support decision, to improve performance) forschool-leader appraisal, it is rarely feasible to do so in practice. School-leadersshould come to understand that although needs for accountability and for data to

support personnel decisions exist, the central purpose for evaluation is perform-

ance improvement. Even in selection situations, where a decision clearly hangs in

the balance, it is important to specify and make known that the process isoriented toward professional development. The selection processes should, infact, be a learning experience for candidates regardless of their success; forunsuccessful candidates the process should direct them to avenues in which they

can better prepare themselves for future selection. For the most part, it should

also support them emotionally so that they remain committed to further applica-

tion.Base criteria on a coherent conception of growth:

Knowledge about what it means to be an expert school-leader is becomingreasonably robust and offers the most promising source of criteria for selection

and appraisal. Conceptions of growth should begin with well specified minimumlevels of expertise but should clarify not only expectations for expert practice but

also manageable steps toward such practice. Expectations for expert practice

26S 241

Developing Expert Leadership for Future Schools

should rely on research-based, practical and professional sources of knowledge,and the goals of the organization. In decision-oriented situations the entire set ofcriteria need to apply to all. Where developmental purposes arc the focus, asubset of criteria should be selected. These should be negotiated between apprais-er and the one to be appraised (in a pre-conference, for example).

Train selectors and appraisers:Those with responsibility for selection and/or appraisal duties should be welltrained in three special domains. First, they should have a clear conception ofgrowth. Second, they should be knowledgeable about local selection and apprais-al procedures, especially including criteria related to expert school leadership.Finally, they should be competent in the skills required to implement the policy.Examples of such skills include, interviewing techniques, observation/data-recording techniques, and inter-personal skills. Responsibilities for school-lcaderselection and appraisal should be clearly outlined in supervisory job descriptionsand incumbents should be held accountable for them.

Selection procedures should be sensitive to motivational characteris-tics of candidates:Perhaps the single best predictor of whether or not principals used informationfrom their own appraisals is hunger for such information. Although, in practice,it is difficult for selection committees to measure motivational qualities validly orreliably, it would be useful for them to look for evidence of candidates' propensi-ty toward self-evaluation, reflection, and an affinity for continuous, professionalgrowth. It seems likely that evidence of such motivation prior to selection wouldpredict, reasonably well, continued motivation of the same type throughout acandidate's tenure as a school-leader.

Data Collection

Selection procedures should make use of 'formal' instruments as earlyscreening devices:Instruments such as the PPBI should be used as early screening mechanisms forselection procedures. In order to ensure fairness, data from self-evaluation andcolleague forms (supervisor, peers) should be submitted for each candidate.District norms should be established to provide a basis for comparison.

Selection procedures should employ 'behavior-description inter-views':'Behavior-description interview' techniques (Jantz et al., 1986) should be adoptedin order to ground candidate responses in actual, past performance. Even ifinappropriate decisions made by respondents are revealed in the interview, theprocess allows for a more fair and accurate basis for decisions (as well as carryingwith it the benefit of stimulating reflection on the part of the candidate). Thctechnique should be guided by questions that are applied to all applicants andgrounded in the criteria identified prior to the process.

Use multiple data sources, collection intervals, and instruments:Valid and reliable data are essential, especially to decision-oriented evaluationsituations. The use of multiple sources of data (e.g., teachers, peers, self-evaluation, supervisor observations, parents, documentation/archives, portfolio,etc.) will greatly enhance the believability of conclusions drawn about indi-



Selection Processes and Measurement Issues

viduals. Comparable information should be available for all candidates andmechanisms for efficient and accurate processing, and summarizing of the datashould be included. Data will need to be collected at different intervals to ensurereliability of 'observations.' In development-oriented appraisals it is not necessary

to be as rigorous in establishing the validity of the data but it is necessary that

findings are viewed as credible by the candidate. Data should be collected on anongoing basis with an appropriate mix of informal and formal methods used.Agreement about data to be collected should be reached by the appraiser andappraisal-subject prior to the growth-oriented process.

Report* and Follow-up

Use multiple selectors/appraisers:Selection dec'sions and other personnel decisions should not be left to single

individuals, in the interest of guarding against susceptibility to real or perceived

political influence and bias. Using a panel of decision makers, especially involv-

ing people from different roles, will help to ensure fairness and the perception of

fairness. This practice is somewhat less necessary in growth-oriented appraisal

situations. Although it will likely be advantageous to the principal being

appraised, the development of credible and trusting relationships between indi-

viduals is more important.Appraisers should communicate with school-leaders continuously:

Just 'as an appraiser dropping in for a 'quick scan' to justify his or her impending

appraisal report can have detrimental effects frequent, albeit brief, visits or phone

calls can have very positive efkcts on school-leaders' propensity to take appraisalinformation seriously. The focus for conversation should be pertinent to the

school-leader's objectives but should bc timely and connected to current concerns

or issues. The conversation should not only be viewed as a progress report but as

an opportunity to share prior experience and knowledge.There is a danger that too much communication may pose a threat to a

school-leader's sense of autonomy. With each case, an appropriate balance be-

tween too much and too little interaction should be sought. School-leaders are

more likely to make appraisers aware that they are overdoing it than they are to

be openly critical about inattention. On the other hand, lack of attention can lead

to cynicism, an outcome that can erode the trust and credibility required for

appraisal information to be used. Developing trust and credibility arc important

objectives.School-leaders should be encouraged to participate in the appraisal

process:School-leaders who take the time and effort to add to the appraiser's knowledge

by compiling different sorts of evidence about themselves (e.g., behavioral

check-lists gathered anonymously from staff, self-assessments, school docu-

rnents, 'Ac.) stand to gain significantly more from the process. Not only will

appraisers be better informed when time comes for discussing performance, but

it is likely school-leaders will have engaged in more reflection about their own

performance.Develop plan for growth:

In both decision-oriented and growth-oriented evaluation situations, constructive

feedback should bc provided. The most useful outcome of this feedback is a

2' 0243

Developing Expert Leadership .lor Future Sihoo

negotiated plan for furthcr development. Individuals that arc unsuccessful inselection processes should be debriefed and a plan for growth should be de-veloped. If appropriate and effective criteria are employed, strengths and weak-nesses should be readily identifiable and suggestions for growth should becomeapparent.

Appraisers should strive to include unexpected results in the assess-ment:Appraisals that tell individuals what they already know may serve as a source ofsatisfaction and reward. But appraisals that point out, in a constructive fashion,not only legitimate strengths but also weaknesses, previously not known or fullyunderstood by the school-leader, make a greater contribution to development.Appraisers should endeavor to teach school-leaders something ibnut themselvesevery time the process occurs, while at same time taking great care not to be

Evolution 0./ Policy

Promote local ownership of policy:Key stakeholders should be involved in policy development and revision pro-cesses. For policies to have substantial impact, there must be significant commit-ment to the philosophy, rationale and spirit of them, and direct involvement intheir development by, for example. those responsible for their implementation.Additionally, efforts can be made in mounting the policy to sustain this source oflocal ownership through promotion and advocacy.

Monitor ongoing implementation of policy:The implementation process should be monitored closely until satisfactoryevidence accumulates, suggesting that implementation has taken a form thatapproximates what was intended. From that point, regular, periodic reviews ofthe policy should take place. Of key concern in such monitoring is the extent ofthe policy's impact (defined particularly in terms of utilization). Also of interestin monitoring are the components or features of the policy, as viewed within thccontext of its implementation. that foster impact. It is through such monitoringthat tine tuning can take place. Evidence of impact emerging from such monitor-ing activities will sustain commitment to the processes by those responsible fortheir Unplementation.


Following all of the guidelines outlined in this chapter for improving appraisaland selection processes would be a formidable challenge, not a challenge manyschool districts are likely to consider realistic, in the short run. We appreciate thisand consider the guidelines, rather, as worthwhile directions for the future,directions to be pursued incrementally. It seems important at the outset, how-ever. to acknowledge the size of the task of developing and implementing defen-sible appraisal and selection processes. Quite candidly, we are trying to create aproblem where many believe none exists, apparently. Evidence reviewed in thesechapters, however, suggests that current practices are. in some instances. 'down-


2 7 1

Selection Processes and Measurement Issues

right shocking!' Not only is there a problem but it is a problem about a set of

school-district processes that crucially affect school-leader development. Poorpractices in these areas do not just have 'no effect'. They have a significantnegative effect on the development of future school leadership. For example,where selection processes in a district are perceived to be based on invalidinformation, promising potential candidates often choose not to become actual

candidates a considerable loss to the district and its future schools. When

appraisal information is not used because it is not timely or is provided by an

appraiser who lacks credibiiity, the accumulated result, in time, contributes to the

continuation of school-leader practices that gradually become counter-productivebecause of changed expectations.

The quality of leadership for future schools depmds, in a significant way, on

our willingness to acknowledge 'the problem' and gradually adopt guidelines of

the sort spelled out in this chapter.


1 The category 'appraiser/es aluation-subject dynamics corresponds to the expansion of

our framework for the use of personnel-evaluation informanon duc to findmgs re-

ported in Cousins (1988), Substudy B (see chapter 13).

2 Patton uses the term 'non-misutilization' to indicate justifiable action. We have substi-

tuted purely in the interest of avoiding the double negative.


2 7 n





Chapter 15

Conclusion: Implications forDistrict Leaders

The role of district leaders in developing leaders for future schools has arisenmany times throughout the book; sometimes as an impediment and sometimes asa significant part of the solution. Research reported in Chapters 2 and 4, forexample, suggested quite forcefully that current school-leaders most often seedistrict administrators and the policies and procedures which they manage ashurdles in their efforts to be more effective. The list of hurdles is quite long. Onthe other hand, some district initiatives are especially helpful for some school-leaders. Non-experts find the planning and support of the district especiallyuseful when implementing externally-initiated policies about which they knowlittle. In Chapter 10, we also noted a dramatic 'district effect' in the pattern ofsocialization experiences reported by school-leaders. This research suggested thatwhile many districts may not contribute helpfully to the socialization of theirschool-leaders, targeted and sustained efforts by a district can be of considerableconsequence.

The remainder of this chapter outlines five 'macro-strategies' which webelieve would be helpful for district leaders to use in their efforts to foster thedevelopment of leaders for futurc schools. Several of these macro-strategiesincorporate, and then further extend, specific implications for district leadershipidentified in earlier chapters. The first three macro-strategies are interdependent;they address similar aspects of the district from subtly but usefully differentperspecti ves.

Design School-District Work as a Curriculum forSchool-Leaders

Chapters 11 and 12 were devoted to school-leadership development throughformai 'instruction' of several sorts. Evidence from Chapter 10 about the range ofexperiences which contribute to the socia'ization of school-leaders, however,placed the contribution of such instruction in perspective. Under thc best ofcircumstances, formal programs contribute importantly but modestly to school-leaders' development. More significant contributions are made by the myriad ofekperienrcs naturally created by school-leaders' work. Because this work is, bydcnninon, 'authentic' and 'situated' (the importance of which was discussed in



Conclusion: Implications for District Leaders

Chapter 12), its impact on thP development of school-leaders is paramount.Indeed, the tasks encountered by school-leaders in their schools and districtsco